i=Tappro
several clients config fixed

b=16771

Fixed incompatible pointer cause by previous commit.

b=16771
i=adilger, johann

Use fix size timestamp instead of time_t in the cfg_marker and keep
compitibility for old cfg_marker on 32bit system.

b=16929

make it easier to read.

b=16929

quench the "defined but not called" compiler warnings.

o=adilger
i=johann
i=tom.wang (wangdi)

Disable file join on production releases.

b=12787
i=johann
i=oleg.drokin

i=Robert.Read
i=Nathan.Rutman
test_127 fix

i=johann (att 19293)
i=nathan (att 19334)
i=jc.lafoucriere
att 19293: locking fixes, error condition checks, remove alloc/free in
qos_calc_rr
att 19334: remove code duplication between llapi_file_create{,_pool};
deprecate positional setstripe params

i=Adilger
get rid of LPROC

b=15966
o=komal.nawandar
i=adilger
i=zhenyu.xu (bobijam)

Description: Re-disable certain /proc logging
Details    : Enable and disable client's offset_stats, extents_stats and
             extents_stats_per_process stats logging on the fly.

i=rread
fix tests for low memory

missing b1_6 patch to skip 0 stats

i=jc.lafoucriere
i=johann
attachments 19188, 19189
pools fix missing unlock
qos remove extra putref

r=shadow

- final part of 16776 being already landed to 1.6.x after shadow's inspection.

r=shadow,wangdi

- fixes wrong checking for obd_stopping in recov_thread which prevents recov_thread from sending llcds to
mds in disconnect time because obd_stopping is set before filter_disconnect() is called.

This causes issues like this:

LustreError: 6912:0:(filter.c:2446:filter_disconnect()) error flushing logs to MDS: rc -19

- fix race of handling early reply in sptlrpc.
- port AT api changes from b1_8 (b16972)
b=16999
r=rread
r=wangdi

i=Adilger
cleanup_obdclass () fix; check_mem_leak t-f fn

b=16206
i=deen
i=johann

i=Adilger
force test_127 to fail if no read|write bytes eq 0

change proc entry name srpc.xxx to srpc_xxx to satisfy 'lctl get_param'.
b=14471
r=wangdi

i=Adilger
replace proc with lctl [set|get]_param

b=16953
i=johann, robert.read

Add mount retry support for liblustre.

b=16303
i=nikita
i=johann

Remove old patch.

b=16303
i=nikita
i=johann

Do not drop reference count for the dentry from VFS when lookup,
VFS will do that by itself.

format fixes in commit 16717

b=17024
i=shadow
i=bobijam

Description: Lustre GPF in {:ptlrpc:ptlrpc_server_free_request+373}
Details    : In case of memory pressure, list_del() can be called twice on
	     req->rq_history_list, causing a kernel oops.

b=15574
i=johann
i=alexey.lyashkov

b=14095
i=nathan.rutman
i=brian

1) add lustre_start utility to start or stop multiple Lustre servers
   from a CSV file
2) fix the utility path issue in bug 14094
3) use 'error_output()' to format error messages
4) move common variables and functions to lc_common
5) fix the "lustre_config -h" issue in bug 14133
6) fix the module_opts issue in bug 14092
7) remove the duplicates from checking lnet connectivity to MGS node

Add full conf log tests for bug 16717.
Fixes for llog_test, and readd it to sanity.
b=16717
i=Nathan,Umka

b=16788
i=adilger
i=nathan.rutman

add per-nid statistics on HEAD.

i=nathan
i=adilger
OST pools on HEAD, comprehensive patch including 17054:19007;
16935:18918,19012,19089,19128; 16978:18872

Branch b_release_1_6_6
b=16303
i=adilger
Description: Oops in statahead when CONFIG_SLUB is enabled
Details    : Disable statahead by default on systems with CONFIG_SLUB enabled,
             until such a time as problem is understood.

r=adilger,robert

- fixes server side scalability issue which became visible with lru resize work landed.
It replcaes linear list of held locks on server with hash table from class_hash.c which improves lock find time drastically for case when one client holds ~limit of server locks (quite a big number for big servers) which is very possible if all clients are idle and one compiles kernel, etc.

b=2066
i=adilger
i=robert.read

b=17043
o=nikita.danilov
i=adilger
i=zhenyu.xu (bobijam)

mount.lustre: match both host name and nid against mount table.

ones.

b=11301
i=johann
i=alexey.lyashkov

b=17023
o=adilger
i=zhenyu.xu (bobijam)
i=johann

per-nid statistics are collected incorrectly.