port Adaptive Timeouts from b1_6 to HEAD.
b=14071
r=nathan
r=adilger

b=14480
i=green

add doxygen doc for ldlm_flock_waitq and ldlm_flock_waitq_lock.

sync procfs/llog_thread with import destory
b=15684
i=Robert, Jay

b=15940
i=isaac
i=oleg

Convert kp30.h to an internal libcfs include and change users of kp30.h to
include libcfs.h if they were not already.

b=14010
i=shadow
i=zam

Short read patch.

fix KEY_IS macro

Branch HEAD
b=14032
i=umka
i=deen

starve other threads from the resource lock for a long time in
case of long waiting queue, so change the debug level from
D_OTHER to the less frequently used D_INFO.

b=15953
i=adilger
i=nathan.rutman

Allow both "clear" and "clear\n" to be written to ldlm.namespaces.*.lru_size.
Minor line wrapping and whitespace cleanup.
b=13868

b=15924
i=green, johann

Severity   : major
Bugzilla   : 15924
Description: do not process already freed flock
Details    : flock can possibly be freed by another thread before it reaches
             to ldlm_flock_completion_ast.

b=14480
o=green
i=shadow, bobijam

Description: LBUG during stress test
Details    : Need properly lock accesses the flock deadlock detection list.

r=nikita,shadow
- committed missed part of 15226. Cleanups in comments, use doxygen style.

b=15776
r=shadow, vitaly

Prevent locks stalls on glimpse races

this allows us to quickly find what pages to write and remove
once lock callback is received.

b=10718
i=green
i=johann

r=shadow,nikita
- fixes issue with accessing freed @ns on behalf rpc handling when some late rpc comes in fini time;
- add new comments in functions, converts existing ones to be inline wigth doxygen comments policies.

- missed part of 15440. Do not forget to call ldlm_pool_fini() before removing ns proc dir;

- call ldlm_namespace_unregister() namespace _before_ __ldlm_namespace_free() which is waiting for refc == 0 as ldlm_namespace_unregister() drops refc.

r=shadow, nikita

- fixes missed list init in CFS_LIST_INIT_HEAD(&ns->ns_list_chain);
- cleanups about registering/unregistering @ns in its list.

b=13310
o=eeb
i=joahnn, nikita

The bug is that target_send_reply_msg() calls ptlrpc_error() if 'rc != 0', but
"difficult" replies may not be passed to rpc_error().

i=green,shadow,vitaly

skiplists are used to group compatible locks on granted list
that was implemented as tracking first and last lock of each lock group
the patch changes that to using doubly linked circular lists

b=11777
r=green, shadow

invalidate the import to halt request handling so the namespace can be safely
cleaned up

b=15851
i=zam

b=14523
i=adilger
i=eric.mei

b=14929
i=zam
i=nikita

b=14034
r=adilger, bobijam

error handling for the ENOMEM case when failing to allocate ldlm resource

b=14929
i=zam

b=14949
i=johan
i=deen

i=vitaly.vertman
i=oleg.drokin

Lockless i/o and lockless truncate code and sanityN tests.

r=Nathan Rutman, Alexey Lyashkov

Ported the patch 'to organize the stats on per-nid basis instead of the existing per-export basis' to HEAD.
This patch also includes all its related bug fixes(bug 14016, bug 14370, bug 14873).

b=14929
i=zam

r=shadow,adilger
- remove @ns from list using list_del_init() to makes sure that consequent list_empty() will really work right on removed list_head.

b=14910
i=umka
i=tappro

update lock on lookup for aged files

b=14661
i=green
i=shadow

-- remove BL_AST lock from lru once we mark the lock as a such;
-- do not place BL_AST lock into lru even if the lock is matched and put later;
   => no BL_AST locks in lru at all;
-- CANCELING locks are not so numerous, but to avoid its traversal by another
   thread another time, drop them from the unused list once come them across.

b=14257

Do not include LdLM_FLOCK locks into ldlm_pool sanity calculations.

b=14533

a spontaneous change got into the commit and prevented HEAD to unmount

r=vitaly,adilger

Always walk entire list of granted locks on match for now.

obd objects and this also block reconnect with -EALREADY.
This isn't fix all lprocfs bugs - but make it rare.
b=14533
i=adilger
i=nathan

ptlrpc connect with V2 message format.

b=13805
i=ericm
i=bobijam

add support for several checksum algorithms. Currently, CRC32 and
Adler-32 are supported. The checksum type can be changed on the fly
through /proc/fs/lustre/osc/*/checksum_type.

Forget to initialize the li_group.

support other encrypt/checksum algorithms for bulk i/o, default in gss
mode is sha1/aes128, for others default enable bulk checksum by adler32.
b=13667
r=yong.fan
r=h.huang
r=adilger

b=11300
r=adilger,oleg,nikita,vitaly

Using interval tree manages the extent locks to accelerate the evicting of contended locks.