i=Adilger

Wait recovery complete before AFTERUSED is checked.

i=Scjody

set stripe count over all OSTs; calculate SIZE based on min kbytesfree.

i=Adilger
i=Nathan

do not create recursive dirs

i=Nathan

remove 18 from A_E list

i=Adilger
i=Nathan

test_18 MDS_MKFS_OPTS fix; cfg/local.sh MKFSOPT fix.

i=Adilger

make ALWAYS_EXCEPT usage more consistent.

new 1.4 image from Andreas.

i=Adilger
b=13649
i=Nathan

SAVE_PWD fix; 32a, 32b 35 fixes: disk1_4 correct path, correct FSNAME.

. "false conflicts" may cause file contention and switching to lockless i/o
. compilation warning on a 64-bit platform
. contention reports are recorded as lustre errors
. llap_from_page complains about wrong llap_checksums for lockless pages
. assertions needed to be sure that ll_releasepage is called on pages with
  locks only
. NS_DEFAULT_CONTENDED_LOCKS=0 was considered as wrong default value

and, the lockless read/write is disabled by default now.

b=11270
i=green
i=adilger

b=13358
i=nikita

add regression test.

b=13358
i=nikita
i=bobijam

Don't LASSERT correctness of data obtained through the network.

i=Adilger

export DIR2 tdir

Add OBD_CONNECT_MDS_MDS and OBD_CONNECT_REAL from HEAD so we don't get any
conflicts in the future.

Add OBD_CONNECT_MDS_MDS and OBD_CONNECT_REAL from HEAD so we don't get any
conflicts in the future.

Fix typo in last commit.

i=andreas, johann

The changed prototype of ext_prepare_callback will be used in lustre if HAVE_EXT_PREPARE_CB_EXTENT is declared.

i=adilger
df after clearing fail_loc

Disable client page cache checksumming by default. This does not affect
the wire checksumming, and only adds a lot of CPU overhead that is mostly
useful only for debugging rare client memory corruption vs. the more
common network transfer corruption. It can be enabled at runtime via
"echo 1 > $LPROC/llite/*/checksum_pages".
b=13579
i=scjody
i=nathan

b=13547, b=13627
Description: Data corruption for OSTs that are formatted larger than 2TB
             on 32-bit servers.
Details    : When generating the bio request for lustre file writes the
	     sector number would overflow a temporary variable before being
	     used for the IO.  The data reads correctly from Lustre (which
	     will overflow in a similar manner) but other file data or
	     filesystem metadata may be corrupted in some cases.

i=Nathan

get rid of "./"

b=13304

b=13457
i=nathan
i=brian

fixes option syntax for lfs

b=13563
i=adilger
i=green

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.

r=nathan,adilger,shadow,green

- separates client and server namespaces. Each "side" has own list and own lock;
- separate pool shrinker to client and server shrinkers which work each with own list. This is needed to avoid mixing up server and client pool cached resources which are too different. Client's locks may be canceled in sync manner and we can return to VM number of still cached resources. And server resources (locks) are not removed in sync way,  we just change SLV and expect that client will cancel something. To VM we return 0 as number of canceled locks;

- in ldlm_pools_shrink() use down_trylock() to avoid locking ns sem when it is already locked. This fixes hang up in test 116 if memory pressure comes. This issue is due to deadlock bewteen shrinker and pool thread if client and server run on same host;

- move lru add stuff into separate func;
- change l_last_used and move lock to tail of lru for case of FL_TEST_LOCK to make sure that it will still hang for some time in lru afer that. So that, if we looked for look even with FL_TEST_LOCK this means that we may need its resourse yet some time and better to stay lock in cache.

b=13358
i=wangdi
i=nikita

Impossible (EPROTO is returned) to access a directory that has a
non-default striping and ACLs.

b=13181
i=alex
i=nikita

Don't take the BKL in fsfilt_ext3_setattr() for 2.6 kernels.
It causes scheduling issues when removing large files.

i=Adilger

Skip "cd ." part for patchless.

- grant_speed, grant_rate, garnt_plan, cancel_rate have own procfs functions which take into account possible schedule lag for pools thread.

b=13304
i=johann
i=green

b=13596
i=adilger
i=deen
original patch by green

- fixed issue with long time not being scheduled pools thread. We need to compensate that in lprocfs_rd_pool_state()

Use correct structure name.

b=13600
i=kalpak, bobijam

original patch by adilger.

Description: "lfs find -obd UUID" prints directories
Details    : "lfs find -obd UUID" will return all directory names instead
             of just file names. It is incorrect because the directories
             do not reside on the OSTs.

b=2369
i=johann, deen

replace inode->i_size access with i_size_read/write

Remove client patches from SLES 10 kernel.  This causes SLES 10 clients to
behave as patchless clients even on a Lustre-patched (server) kernel.

b=12411
i=adilger
i=johann

i=Nathan

Increase a set of files to choose from.

Move bug 13547 to 1.6.3 section instead of 1.6.4

b=12186
Update ChangeLog

b=12186
i=Adilger

Fixes man pages for lfs