8 years agomm: strictly nested kmap_atomic()
Peter Zijlstra [Tue, 26 Oct 2010 21:21:47 +0000]
mm: strictly nested kmap_atomic()

Ensure kmap_atomic() usage is strictly nested

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Reviewed-by: Rik van Riel <riel@redhat.com>
Acked-by: Chris Metcalf <cmetcalf@tilera.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: David Miller <davem@davemloft.net>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan,tmpfs: treat used once pages on tmpfs as used once
KOSAKI Motohiro [Tue, 26 Oct 2010 21:21:46 +0000]
vmscan,tmpfs: treat used once pages on tmpfs as used once

When a page has PG_referenced, shrink_page_list() discards it only if it
is not dirty.  This rule works fine if the backing filesystem is a regular
one.  PG_dirty is a good signal that the page was used recently because
the flusher threads clean pages periodically.  In addition, page writeback
is costlier than simple page discard.

However, when a page is on tmpfs this heuristic doesn't work because
flusher threads don't write back tmpfs pages.  Consequently tmpfs pages
always rotate around the lru twice at least and adds unnecessary lru
churn.  Simple tmpfs streaming io shouldn't cause large anonymous page
swap-out.

Remove this unncessary reclaim bonus of tmpfs pages.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Hugh Dickins <hughd@google.com>
Reviewed-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: remove the internal 5% low bound on dirty_ratio
Wu Fengguang [Tue, 26 Oct 2010 21:21:45 +0000]
writeback: remove the internal 5% low bound on dirty_ratio

The dirty_ratio was silently limited in global_dirty_limits() to >= 5%.
This is not a user expected behavior.  And it's inconsistent with
calc_period_shift(), which uses the plain vm_dirty_ratio value.

Let's remove the internal bound.

At the same time, fix balance_dirty_pages() to work with the
dirty_thresh=0 case.  This allows applications to proceed when
dirty+writeback pages are all cleaned.

And ">" fits with the name "exceeded" better than ">=" does.  Neil thinks
it is an aesthetic improvement as well as a functional one :)

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Jan Kara <jack@suse.cz>
Proposed-by: Con Kolivas <kernel@kolivas.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Michael Rubin <mrubin@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: do not sleep on the congestion queue if there are no congested BDIs or...
Mel Gorman [Tue, 26 Oct 2010 21:21:45 +0000]
writeback: do not sleep on the congestion queue if there are no congested BDIs or if significant congestion is not being encountered in the current zone

If congestion_wait() is called with no BDI congested, the caller will
sleep for the full timeout and this may be an unnecessary sleep.  This
patch adds a wait_iff_congested() that checks congestion and only sleeps
if a BDI is congested else, it calls cond_resched() to ensure the caller
is not hogging the CPU longer than its quota but otherwise will not sleep.

This is aimed at reducing some of the major desktop stalls reported during
IO.  For example, while kswapd is operating, it calls congestion_wait()
but it could just have been reclaiming clean page cache pages with no
congestion.  Without this patch, it would sleep for a full timeout but
after this patch, it'll just call schedule() if it has been on the CPU too
long.  Similar logic applies to direct reclaimers that are not making
enough progress.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan: isolate_lru_pages(): stop neighbour search if neighbour cannot be isolated
KOSAKI Motohiro [Tue, 26 Oct 2010 21:21:44 +0000]
vmscan: isolate_lru_pages(): stop neighbour search if neighbour cannot be isolated

isolate_lru_pages() does not just isolate LRU tail pages, but also
isolates neighbour pages of the eviction page.  The neighbour search does
not stop even if neighbours cannot be isolated which is excessive as the
lumpy reclaim will no longer result in a successful higher order
allocation.  This patch stops the PFN neighbour pages if an isolation
fails and moves on to the next block.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan: remove dead code in shrink_inactive_list()
KOSAKI Motohiro [Tue, 26 Oct 2010 21:21:43 +0000]
vmscan: remove dead code in shrink_inactive_list()

After synchrounous lumpy reclaim, the page_list is guaranteed to not have
active pages as page activation in shrink_page_list() disables lumpy
reclaim.  Remove the dead code.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan: narrow the scenarios in whcih lumpy reclaim uses synchrounous reclaim
KOSAKI Motohiro [Tue, 26 Oct 2010 21:21:42 +0000]
vmscan: narrow the scenarios in whcih lumpy reclaim uses synchrounous reclaim

shrink_page_list() can decide to give up reclaiming a page under a
number of conditions such as

  1. trylock_page() failure
  2. page is unevictable
  3. zone reclaim and page is mapped
  4. PageWriteback() is true
  5. page is swapbacked and swap is full
  6. add_to_swap() failure
  7. page is dirty and gfpmask don't have GFP_IO, GFP_FS
  8. page is pinned
  9. IO queue is congested
 10. pageout() start IO, but not finished

With lumpy reclaim, failures result in entering synchronous lumpy reclaim
but this can be unnecessary.  In cases (2), (3), (5), (6), (7) and (8),
there is no point retrying.  This patch causes lumpy reclaim to abort when
it is known it will fail.

Case (9) is more interesting. current behavior is,
  1. start shrink_page_list(async)
  2. found queue_congested()
  3. skip pageout write
  4. still start shrink_page_list(sync)
  5. wait on a lot of pages
  6. again, found queue_congested()
  7. give up pageout write again

So, it's useless time wasting.  However, just skipping page reclaim is
also notgood as x86 allocating a huge page needs 512 pages for example.
It can have more dirty pages than queue congestion threshold (~=128).

After this patch, pageout() behaves as follows;

 - If order > PAGE_ALLOC_COSTLY_ORDER
Ignore queue congestion always.
 - If order <= PAGE_ALLOC_COSTLY_ORDER
skip write page and disable lumpy reclaim.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan: synchronous lumpy reclaim should not call congestion_wait()
KOSAKI Motohiro [Tue, 26 Oct 2010 21:21:41 +0000]
vmscan: synchronous lumpy reclaim should not call congestion_wait()

congestion_wait() means "wait until queue congestion is cleared".
However, synchronous lumpy reclaim does not need this congestion_wait() as
shrink_page_list(PAGEOUT_IO_SYNC) uses wait_on_page_writeback() and it
provides the necessary waiting.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: account for time spent congestion_waited
Mel Gorman [Tue, 26 Oct 2010 21:21:41 +0000]
writeback: account for time spent congestion_waited

There is strong evidence to indicate a lot of time is being spent in
congestion_wait(), some of it unnecessarily.  This patch adds a tracepoint
for congestion_wait to record when congestion_wait() was called, how long
the timeout was for and how long it actually slept.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agotracing, vmscan: add trace events for LRU list shrinking
Mel Gorman [Tue, 26 Oct 2010 21:21:40 +0000]
tracing, vmscan: add trace events for LRU list shrinking

There have been numerous reports of stalls that pointed at the problem
being somewhere in the VM.  There are multiple roots to the problems which
means dealing with any of the root problems in isolation is tricky to
justify on their own and they would still need integration testing.  This
patch series puts together two different patch sets which in combination
should tackle some of the root causes of latency problems being reported.

Patch 1 adds a tracepoint for shrink_inactive_list.  For this series, the
most important results is being able to calculate the scanning/reclaim
ratio as a measure of the amount of work being done by page reclaim.

Patch 2 accounts for time spent in congestion_wait.

Patches 3-6 were originally developed by Kosaki Motohiro but reworked for
this series.  It has been noted that lumpy reclaim is far too aggressive
and trashes the system somewhat.  As SLUB uses high-order allocations, a
large cost incurred by lumpy reclaim will be noticeable.  It was also
reported during transparent hugepage support testing that lumpy reclaim
was trashing the system and these patches should mitigate that problem
without disabling lumpy reclaim.

Patch 7 adds wait_iff_congested() and replaces some callers of
congestion_wait().  wait_iff_congested() only sleeps if there is a BDI
that is currently congested.  Patch 8 notes that any BDI being congested
is not necessarily a problem because there could be multiple BDIs of
varying speeds and numberous zones.  It attempts to track when a zone
being reclaimed contains many pages backed by a congested BDI and if so,
reclaimers wait on the congestion queue.

I ran a number of tests with monitoring on X86, X86-64 and PPC64. Each
machine had 3G of RAM and the CPUs were

X86:    Intel P4 2-core
X86-64: AMD Phenom 4-core
PPC64:  PPC970MP

Each used a single disk and the onboard IO controller.  Dirty ratio was
left at 20.  I'm just going to report for X86-64 and PPC64 in a vague
attempt to keep this report short.  Four kernels were tested each based on
v2.6.36-rc4

traceonly-v2r2:     Patches 1 and 2 to instrument vmscan reclaims and congestion_wait
lowlumpy-v2r3:      Patches 1-6 to test if lumpy reclaim is better
waitcongest-v2r3:   Patches 1-7 to only wait on congestion
waitwriteback-v2r4: Patches 1-8 to detect when a zone is congested

nocongest-v1r5: Patches 1-3 for testing wait_iff_congestion
nodirect-v1r5:  Patches 1-10 to disable filesystem writeback for better IO

The tests run were as follows

kernbench
compile-based benchmark. Smoke test performance

sysbench
OLTP read-only benchmark. Will be re-run in the future as read-write

micro-mapped-file-stream
This is a micro-benchmark from Johannes Weiner that accesses a
large sparse-file through mmap(). It was configured to run in only
single-CPU mode but can be indicative of how well page reclaim
identifies suitable pages.

stress-highalloc
Tries to allocate huge pages under heavy load.

kernbench, iozone and sysbench did not report any performance regression
on any machine.  sysbench did pressure the system lightly and there was
reclaim activity but there were no difference of major interest between
the kernels.

X86-64 micro-mapped-file-stream

                                      traceonly-v2r2           lowlumpy-v2r3        waitcongest-v2r3     waitwriteback-v2r4
pgalloc_dma                       1639.00 (   0.00%)       667.00 (-145.73%)      1167.00 ( -40.45%)       578.00 (-183.56%)
pgalloc_dma32                  2842410.00 (   0.00%)   2842626.00 (   0.01%)   2843043.00 (   0.02%)   2843014.00 (   0.02%)
pgalloc_normal                       0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pgsteal_dma                        729.00 (   0.00%)        85.00 (-757.65%)       609.00 ( -19.70%)       125.00 (-483.20%)
pgsteal_dma32                  2338721.00 (   0.00%)   2447354.00 (   4.44%)   2429536.00 (   3.74%)   2436772.00 (   4.02%)
pgsteal_normal                       0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pgscan_kswapd_dma                 1469.00 (   0.00%)       532.00 (-176.13%)      1078.00 ( -36.27%)       220.00 (-567.73%)
pgscan_kswapd_dma32            4597713.00 (   0.00%)   4503597.00 (  -2.09%)   4295673.00 (  -7.03%)   3891686.00 ( -18.14%)
pgscan_kswapd_normal                 0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pgscan_direct_dma                   71.00 (   0.00%)       134.00 (  47.01%)       243.00 (  70.78%)       352.00 (  79.83%)
pgscan_direct_dma32             305820.00 (   0.00%)    280204.00 (  -9.14%)    600518.00 (  49.07%)    957485.00 (  68.06%)
pgscan_direct_normal                 0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pageoutrun                       16296.00 (   0.00%)     21254.00 (  23.33%)     18447.00 (  11.66%)     20067.00 (  18.79%)
allocstall                         443.00 (   0.00%)       273.00 ( -62.27%)       513.00 (  13.65%)      1568.00 (  71.75%)

These are based on the raw figures taken from /proc/vmstat.  It's a rough
measure of reclaim activity.  Note that allocstall counts are higher
because we are entering direct reclaim more often as a result of not
sleeping in congestion.  In itself, it's not necessarily a bad thing.
It's easier to get a view of what happened from the vmscan tracepoint
report.

FTrace Reclaim Statistics: vmscan

                                traceonly-v2r2   lowlumpy-v2r3 waitcongest-v2r3 waitwriteback-v2r4
Direct reclaims                                443        273        513       1568
Direct reclaim pages scanned                305968     280402     600825     957933
Direct reclaim pages reclaimed               43503      19005      30327     117191
Direct reclaim write file async I/O              0          0          0          0
Direct reclaim write anon async I/O              0          3          4         12
Direct reclaim write file sync I/O               0          0          0          0
Direct reclaim write anon sync I/O               0          0          0          0
Wake kswapd requests                        187649     132338     191695     267701
Kswapd wakeups                                   3          1          4          1
Kswapd pages scanned                       4599269    4454162    4296815    3891906
Kswapd pages reclaimed                     2295947    2428434    2399818    2319706
Kswapd reclaim write file async I/O              1          0          1          1
Kswapd reclaim write anon async I/O             59        187         41        222
Kswapd reclaim write file sync I/O               0          0          0          0
Kswapd reclaim write anon sync I/O               0          0          0          0
Time stalled direct reclaim (seconds)         4.34       2.52       6.63       2.96
Time kswapd awake (seconds)                  11.15      10.25      11.01      10.19

Total pages scanned                        4905237   4734564   4897640   4849839
Total pages reclaimed                      2339450   2447439   2430145   2436897
%age total pages scanned/reclaimed          47.69%    51.69%    49.62%    50.25%
%age total pages scanned/written             0.00%     0.00%     0.00%     0.00%
%age  file pages scanned/written             0.00%     0.00%     0.00%     0.00%
Percentage Time Spent Direct Reclaim        29.23%    19.02%    38.48%    20.25%
Percentage Time kswapd Awake                78.58%    78.85%    76.83%    79.86%

What is interesting here for nocongest in particular is that while direct
reclaim scans more pages, the overall number of pages scanned remains the
same and the ratio of pages scanned to pages reclaimed is more or less the
same.  In other words, while we are sleeping less, reclaim is not doing
more work and as direct reclaim and kswapd is awake for less time, it
would appear to be doing less work.

FTrace Reclaim Statistics: congestion_wait
Direct number congest     waited                87        196         64          0
Direct time   congest     waited            4604ms     4732ms     5420ms        0ms
Direct full   congest     waited                72        145         53          0
Direct number conditional waited                 0          0        324       1315
Direct time   conditional waited               0ms        0ms        0ms        0ms
Direct full   conditional waited                 0          0          0          0
KSwapd number congest     waited                20         10         15          7
KSwapd time   congest     waited            1264ms      536ms      884ms      284ms
KSwapd full   congest     waited                10          4          6          2
KSwapd number conditional waited                 0          0          0          0
KSwapd time   conditional waited               0ms        0ms        0ms        0ms
KSwapd full   conditional waited                 0          0          0          0

The vanilla kernel spent 8 seconds asleep in direct reclaim and no time at
all asleep with the patches.

MMTests Statistics: duration
User/Sys Time Running Test (seconds)         10.51     10.73      10.6     11.66
Total Elapsed Time (seconds)                 14.19     13.00     14.33     12.76

Overall, the tests completed faster. It is interesting to note that backing off further
when a zone is congested and not just a BDI was more efficient overall.

PPC64 micro-mapped-file-stream
pgalloc_dma                    3024660.00 (   0.00%)   3027185.00 (   0.08%)   3025845.00 (   0.04%)   3026281.00 (   0.05%)
pgalloc_normal                       0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pgsteal_dma                    2508073.00 (   0.00%)   2565351.00 (   2.23%)   2463577.00 (  -1.81%)   2532263.00 (   0.96%)
pgsteal_normal                       0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pgscan_kswapd_dma              4601307.00 (   0.00%)   4128076.00 ( -11.46%)   3912317.00 ( -17.61%)   3377165.00 ( -36.25%)
pgscan_kswapd_normal                 0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pgscan_direct_dma               629825.00 (   0.00%)    971622.00 (  35.18%)   1063938.00 (  40.80%)   1711935.00 (  63.21%)
pgscan_direct_normal                 0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)         0.00 (   0.00%)
pageoutrun                       27776.00 (   0.00%)     20458.00 ( -35.77%)     18763.00 ( -48.04%)     18157.00 ( -52.98%)
allocstall                         977.00 (   0.00%)      2751.00 (  64.49%)      2098.00 (  53.43%)      5136.00 (  80.98%)

Similar trends to x86-64. allocstalls are up but it's not necessarily bad.

FTrace Reclaim Statistics: vmscan
Direct reclaims                                977       2709       2098       5136
Direct reclaim pages scanned                629825     963814    1063938    1711935
Direct reclaim pages reclaimed               75550     242538     150904     387647
Direct reclaim write file async I/O              0          0          0          2
Direct reclaim write anon async I/O              0         10          0          4
Direct reclaim write file sync I/O               0          0          0          0
Direct reclaim write anon sync I/O               0          0          0          0
Wake kswapd requests                        392119    1201712     571935     571921
Kswapd wakeups                                   3          2          3          3
Kswapd pages scanned                       4601307    4128076    3912317    3377165
Kswapd pages reclaimed                     2432523    2318797    2312673    2144616
Kswapd reclaim write file async I/O             20          1          1          1
Kswapd reclaim write anon async I/O             57        132         11        121
Kswapd reclaim write file sync I/O               0          0          0          0
Kswapd reclaim write anon sync I/O               0          0          0          0
Time stalled direct reclaim (seconds)         6.19       7.30      13.04      10.88
Time kswapd awake (seconds)                  21.73      26.51      25.55      23.90

Total pages scanned                        5231132   5091890   4976255   5089100
Total pages reclaimed                      2508073   2561335   2463577   2532263
%age total pages scanned/reclaimed          47.95%    50.30%    49.51%    49.76%
%age total pages scanned/written             0.00%     0.00%     0.00%     0.00%
%age  file pages scanned/written             0.00%     0.00%     0.00%     0.00%
Percentage Time Spent Direct Reclaim        18.89%    20.65%    32.65%    27.65%
Percentage Time kswapd Awake                72.39%    80.68%    78.21%    77.40%

Again, a similar trend that the congestion_wait changes mean that direct
reclaim scans more pages but the overall number of pages scanned while
slightly reduced, are very similar.  The ratio of scanning/reclaimed
remains roughly similar.  The downside is that kswapd and direct reclaim
was awake longer and for a larger percentage of the overall workload.
It's possible there were big differences in the amount of time spent
reclaiming slab pages between the different kernels which is plausible
considering that the micro tests runs after fsmark and sysbench.

Trace Reclaim Statistics: congestion_wait
Direct number congest     waited               845       1312        104          0
Direct time   congest     waited           19416ms    26560ms     7544ms        0ms
Direct full   congest     waited               745       1105         72          0
Direct number conditional waited                 0          0       1322       2935
Direct time   conditional waited               0ms        0ms       12ms      312ms
Direct full   conditional waited                 0          0          0          3
KSwapd number congest     waited                39        102         75         63
KSwapd time   congest     waited            2484ms     6760ms     5756ms     3716ms
KSwapd full   congest     waited                20         48         46         25
KSwapd number conditional waited                 0          0          0          0
KSwapd time   conditional waited               0ms        0ms        0ms        0ms
KSwapd full   conditional waited                 0          0          0          0

The vanilla kernel spent 20 seconds asleep in direct reclaim and only
312ms asleep with the patches.  The time kswapd spent congest waited was
also reduced by a large factor.

MMTests Statistics: duration
ser/Sys Time Running Test (seconds)         26.58     28.05      26.9     28.47
Total Elapsed Time (seconds)                 30.02     32.86     32.67     30.88

With all patches applies, the completion times are very similar.

X86-64 STRESS-HIGHALLOC
                traceonly-v2r2     lowlumpy-v2r3  waitcongest-v2r3waitwriteback-v2r4
Pass 1          82.00 ( 0.00%)    84.00 ( 2.00%)    85.00 ( 3.00%)    85.00 ( 3.00%)
Pass 2          90.00 ( 0.00%)    87.00 (-3.00%)    88.00 (-2.00%)    89.00 (-1.00%)
At Rest         92.00 ( 0.00%)    90.00 (-2.00%)    90.00 (-2.00%)    91.00 (-1.00%)

Success figures across the board are broadly similar.

                traceonly-v2r2     lowlumpy-v2r3  waitcongest-v2r3waitwriteback-v2r4
Direct reclaims                               1045        944        886        887
Direct reclaim pages scanned                135091     119604     109382     101019
Direct reclaim pages reclaimed               88599      47535      47863      46671
Direct reclaim write file async I/O            494        283        465        280
Direct reclaim write anon async I/O          29357      13710      16656      13462
Direct reclaim write file sync I/O             154          2          2          3
Direct reclaim write anon sync I/O           14594        571        509        561
Wake kswapd requests                          7491        933        872        892
Kswapd wakeups                                 814        778        731        780
Kswapd pages scanned                       7290822   15341158   11916436   13703442
Kswapd pages reclaimed                     3587336    3142496    3094392    3187151
Kswapd reclaim write file async I/O          91975      32317      28022      29628
Kswapd reclaim write anon async I/O        1992022     789307     829745     849769
Kswapd reclaim write file sync I/O               0          0          0          0
Kswapd reclaim write anon sync I/O               0          0          0          0
Time stalled direct reclaim (seconds)      4588.93    2467.16    2495.41    2547.07
Time kswapd awake (seconds)                2497.66    1020.16    1098.06    1176.82

Total pages scanned                        7425913  15460762  12025818  13804461
Total pages reclaimed                      3675935   3190031   3142255   3233822
%age total pages scanned/reclaimed          49.50%    20.63%    26.13%    23.43%
%age total pages scanned/written            28.66%     5.41%     7.28%     6.47%
%age  file pages scanned/written             1.25%     0.21%     0.24%     0.22%
Percentage Time Spent Direct Reclaim        57.33%    42.15%    42.41%    42.99%
Percentage Time kswapd Awake                43.56%    27.87%    29.76%    31.25%

Scanned/reclaimed ratios again look good with big improvements in
efficiency.  The Scanned/written ratios also look much improved.  With a
better scanned/written ration, there is an expectation that IO would be
more efficient and indeed, the time spent in direct reclaim is much
reduced by the full series and kswapd spends a little less time awake.

Overall, indications here are that allocations were happening much faster
and this can be seen with a graph of the latency figures as the
allocations were taking place
http://www.csn.ul.ie/~mel/postings/vmscanreduce-20101509/highalloc-interlatency-hydra-mean.ps

FTrace Reclaim Statistics: congestion_wait
Direct number congest     waited              1333        204        169          4
Direct time   congest     waited           78896ms     8288ms     7260ms      200ms
Direct full   congest     waited               756         92         69          2
Direct number conditional waited                 0          0         26        186
Direct time   conditional waited               0ms        0ms        0ms     2504ms
Direct full   conditional waited                 0          0          0         25
KSwapd number congest     waited                 4        395        227        282
KSwapd time   congest     waited             384ms    25136ms    10508ms    18380ms
KSwapd full   congest     waited                 3        232         98        176
KSwapd number conditional waited                 0          0          0          0
KSwapd time   conditional waited               0ms        0ms        0ms        0ms
KSwapd full   conditional waited                 0          0          0          0
KSwapd full   conditional waited               318          0        312          9

Overall, the time spent speeping is reduced.  kswapd is still hitting
congestion_wait() but that is because there are callers remaining where it
wasn't clear in advance if they should be changed to wait_iff_congested()
or not.  Overall the sleep imes are reduced though - from 79ish seconds to
about 19.

MMTests Statistics: duration
User/Sys Time Running Test (seconds)       3415.43   3386.65   3388.39    3377.5
Total Elapsed Time (seconds)               5733.48   3660.33   3689.41   3765.39

With the full series, the time to complete the tests are reduced by 30%

PPC64 STRESS-HIGHALLOC
                traceonly-v2r2     lowlumpy-v2r3  waitcongest-v2r3waitwriteback-v2r4
Pass 1          17.00 ( 0.00%)    34.00 (17.00%)    38.00 (21.00%)    43.00 (26.00%)
Pass 2          25.00 ( 0.00%)    37.00 (12.00%)    42.00 (17.00%)    46.00 (21.00%)
At Rest         49.00 ( 0.00%)    43.00 (-6.00%)    45.00 (-4.00%)    51.00 ( 2.00%)

Success rates there are *way* up particularly considering that the 16MB
huge pages on PPC64 mean that it's always much harder to allocate them.

FTrace Reclaim Statistics: vmscan
              stress-highalloc  stress-highalloc  stress-highalloc  stress-highalloc
                traceonly-v2r2     lowlumpy-v2r3  waitcongest-v2r3waitwriteback-v2r4
Direct reclaims                                499        505        564        509
Direct reclaim pages scanned                223478      41898      51818      45605
Direct reclaim pages reclaimed              137730      21148      27161      23455
Direct reclaim write file async I/O            399        136        162        136
Direct reclaim write anon async I/O          46977       2865       4686       3998
Direct reclaim write file sync I/O              29          0          1          3
Direct reclaim write anon sync I/O           31023        159        237        239
Wake kswapd requests                           420        351        360        326
Kswapd wakeups                                 185        294        249        277
Kswapd pages scanned                      15703488   16392500   17821724   17598737
Kswapd pages reclaimed                     5808466    2908858    3139386    3145435
Kswapd reclaim write file async I/O         159938      18400      18717      13473
Kswapd reclaim write anon async I/O        3467554     228957     322799     234278
Kswapd reclaim write file sync I/O               0          0          0          0
Kswapd reclaim write anon sync I/O               0          0          0          0
Time stalled direct reclaim (seconds)      9665.35    1707.81    2374.32    1871.23
Time kswapd awake (seconds)                9401.21    1367.86    1951.75    1328.88

Total pages scanned                       15926966  16434398  17873542  17644342
Total pages reclaimed                      5946196   2930006   3166547   3168890
%age total pages scanned/reclaimed          37.33%    17.83%    17.72%    17.96%
%age total pages scanned/written            23.27%     1.52%     1.94%     1.43%
%age  file pages scanned/written             1.01%     0.11%     0.11%     0.08%
Percentage Time Spent Direct Reclaim        44.55%    35.10%    41.42%    36.91%
Percentage Time kswapd Awake                86.71%    43.58%    52.67%    41.14%

While the scanning rates are slightly up, the scanned/reclaimed and
scanned/written figures are much improved.  The time spent in direct
reclaim and with kswapd are massively reduced, mostly by the lowlumpy
patches.

FTrace Reclaim Statistics: congestion_wait
Direct number congest     waited               725        303        126          3
Direct time   congest     waited           45524ms     9180ms     5936ms      300ms
Direct full   congest     waited               487        190         52          3
Direct number conditional waited                 0          0        200        301
Direct time   conditional waited               0ms        0ms        0ms     1904ms
Direct full   conditional waited                 0          0          0         19
KSwapd number congest     waited                 0          2         23          4
KSwapd time   congest     waited               0ms      200ms      420ms      404ms
KSwapd full   congest     waited                 0          2          2          4
KSwapd number conditional waited                 0          0          0          0
KSwapd time   conditional waited               0ms        0ms        0ms        0ms
KSwapd full   conditional waited                 0          0          0          0

Not as dramatic a story here but the time spent asleep is reduced and we
can still see what wait_iff_congested is going to sleep when necessary.

MMTests Statistics: duration
User/Sys Time Running Test (seconds)      12028.09   3157.17   3357.79   3199.16
Total Elapsed Time (seconds)              10842.07   3138.72   3705.54   3229.85

The time to complete this test goes way down.  With the full series, we
are allocating over twice the number of huge pages in 30% of the time and
there is a corresponding impact on the allocation latency graph available
at.

http://www.csn.ul.ie/~mel/postings/vmscanreduce-20101509/highalloc-interlatency-powyah-mean.ps

This patch:

Add a trace event for shrink_inactive_list() and updates the sample
postprocessing script appropriately.  It can be used to determine how many
pages were reclaimed and for non-lumpy reclaim where exactly the pages
were reclaimed from.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan: delete dead code
Shaohua Li [Tue, 26 Oct 2010 21:21:37 +0000]
vmscan: delete dead code

`priority' cannot be negative here.  And the comment is obsolete.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: fix typo in mm.h when NODE_NOT_IN_PAGE_FLAGS
Will Deacon [Tue, 26 Oct 2010 21:21:37 +0000]
mm: fix typo in mm.h when NODE_NOT_IN_PAGE_FLAGS

NODE_NOT_IN_PAGE_FLAGS is defined in mm.h when the node information is not
stored in the page flags bitmap.

Unfortunately, there's a typo in one of the checks for it.  This patch
fixes it (s/NODE_NOT_IN_PAGEFLAGS/NODE_NOT_IN_PAGE_FLAGS/).  Since this
has been around for ages, I doubt it's been causing any serious problems.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: report dirty thresholds in /proc/vmstat
Michael Rubin [Tue, 26 Oct 2010 21:21:36 +0000]
writeback: report dirty thresholds in /proc/vmstat

The kernel already exposes the user desired thresholds in /proc/sys/vm
with dirty_background_ratio and background_ratio.  But the kernel may
alter the number requested without giving the user any indication that is
the case.

Knowing the actual ratios the kernel is honoring can help app developers
understand how their buffered IO will be sent to the disk.

        $ grep threshold /proc/vmstat
        nr_dirty_threshold 409111
        nr_dirty_background_threshold 818223

Signed-off-by: Michael Rubin <mrubin@google.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: add /sys/devices/system/node/<node>/vmstat
Michael Rubin [Tue, 26 Oct 2010 21:21:35 +0000]
writeback: add /sys/devices/system/node/<node>/vmstat

For NUMA node systems it is important to have visibility in memory
characteristics.  Two of the /proc/vmstat values "nr_written" and
"nr_dirtied" are added here.

# cat /sys/devices/system/node/node20/vmstat
nr_written 0
nr_dirtied 0

Signed-off-by: Michael Rubin <mrubin@google.com>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: add nr_dirtied and nr_written to /proc/vmstat
Michael Rubin [Tue, 26 Oct 2010 21:21:35 +0000]
writeback: add nr_dirtied and nr_written to /proc/vmstat

To help developers and applications gain visibility into writeback
behaviour adding two entries to vm_stat_items and /proc/vmstat.  This will
allow us to track the "written" and "dirtied" counts.

   # grep nr_dirtied /proc/vmstat
   nr_dirtied 3747
   # grep nr_written /proc/vmstat
   nr_written 3618

Signed-off-by: Michael Rubin <mrubin@google.com>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: add account_page_writeback()
Michael Rubin [Tue, 26 Oct 2010 21:21:33 +0000]
mm: add account_page_writeback()

To help developers and applications gain visibility into writeback
behaviour this patch adds two counters to /proc/vmstat.

  # grep nr_dirtied /proc/vmstat
  nr_dirtied 3747
  # grep nr_written /proc/vmstat
  nr_written 3618

These entries allow user apps to understand writeback behaviour over time
and learn how it is impacting their performance.  Currently there is no
way to inspect dirty and writeback speed over time.  It's not possible for
nr_dirty/nr_writeback.

These entries are necessary to give visibility into writeback behaviour.
We have /proc/diskstats which lets us understand the io in the block
layer.  We have blktrace for more in depth understanding.  We have
e2fsprogs and debugsfs to give insight into the file systems behaviour,
but we don't offer our users the ability understand what writeback is
doing.  There is no way to know how active it is over the whole system, if
it's falling behind or to quantify it's efforts.  With these values
exported users can easily see how much data applications are sending
through writeback and also at what rates writeback is processing this
data.  Comparing the rates of change between the two allow developers to
see when writeback is not able to keep up with incoming traffic and the
rate of dirty memory being sent to the IO back end.  This allows folks to
understand their io workloads and track kernel issues.  Non kernel
engineers at Google often use these counters to solve puzzling performance
problems.

Patch #4 adds a pernode vmstat file with nr_dirtied and nr_written

Patch #5 add writeback thresholds to /proc/vmstat

Currently these values are in debugfs. But they should be promoted to
/proc since they are useful for developers who are writing databases
and file servers and are not debugging the kernel.

The output is as below:

 # grep threshold /proc/vmstat
 nr_pages_dirty_threshold 409111
 nr_pages_dirty_background_threshold 818223

This patch:

This allows code outside of the mm core to safely manipulate page
writeback state and not worry about the other accounting.  Not using these
routines means that some code will lose track of the accounting and we get
bugs.

Modify nilfs2 to use interface.

Signed-off-by: Michael Rubin <mrubin@google.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp>
Cc: Jiro SEKIBA <jir@unicus.jp>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm/mempolicy.c: check return code of check_range
Vasiliy Kulikov [Tue, 26 Oct 2010 21:21:32 +0000]
mm/mempolicy.c: check return code of check_range

Function check_range may return ERR_PTR(...). Check for it.

Signed-off-by: Vasiliy Kulikov <segooon@gmail.com>
Acked-by: David Rientjes <rientjes@google.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmscan: prevent background aging of anon page in no swap system
Minchan Kim [Tue, 26 Oct 2010 21:21:31 +0000]
vmscan: prevent background aging of anon page in no swap system

Ying Han reported that backing aging of anon pages in no swap system
causes unnecessary TLB flush.

When I sent a patch(69c8548175), I wanted this patch but Rik pointed out
and allowed aging of anon pages to give a chance to promote from inactive
to active LRU.

It has a two problem.

1) non-swap system

Never make sense to age anon pages.

2) swap configured but still doesn't swapon

It doesn't make sense to age anon pages until swap-on time.  But it's
arguable.  If we have aged anon pages by swapon, VM have moved anon pages
from active to inactive.  And in the time swapon by admin, the VM can't
reclaim hot pages so we can protect hot pages swapout.

But let's think about it.  When does swap-on happen?  It depends on admin.
 we can't expect it.  Nonetheless, we have done aging of anon pages to
protect hot pages swapout.  It means we lost run time overhead when below
high watermark but gain hot page swap-[in/out] overhead when VM decide
swapout.  Is it true?  Let's think more detail.  We don't promote anon
pages in case of non-swap system.  So even though VM does aging of anon
pages, the pages would be in inactive LRU for a long time.  It means many
of pages in there would mark access bit again.  So access bit hot/code
separation would be pointless.

This patch prevents unnecessary anon pages demotion in not-yet-swapon and
non-configured swap system.  Even, in non-configuared swap system
inactive_anon_is_low can be compiled out.

It could make side effect that hot anon pages could swap out when admin
does swap on.  But I think sooner or later it would be steady state.  So
it's not a big problem.

We could lose someting but gain more thing(TLB flush and unnecessary
function call to demote anon pages).

Signed-off-by: Ying Han <yinghan@google.com>
Signed-off-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomemory hotplug: unify is_removable and offline detection code
KAMEZAWA Hiroyuki [Tue, 26 Oct 2010 21:21:30 +0000]
memory hotplug: unify is_removable and offline detection code

Now, sysfs interface of memory hotplug shows whether the section is
removable or not.  But it checks only migrateype of pages and doesn't
check details of cluster of pages.

Next, memory hotplug's set_migratetype_isolate() has the same kind of
check, too.

This patch adds the function __count_unmovable_pages() and makes above 2
checks to use the same logic.  Then, is_removable and hotremove code uses
the same logic.  No changes in the hotremove logic itself.

TODO: need to find a way to check RECLAMABLE. But, considering bit,
      calling shrink_slab() against a range before starting memory hotremove
      sounds better. If so, this patch's logic doesn't need to be changed.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reported-by: Michal Hocko <mhocko@suse.cz>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomemory hotplug: fix notifier's return value check
KAMEZAWA Hiroyuki [Tue, 26 Oct 2010 21:21:29 +0000]
memory hotplug: fix notifier's return value check

Even if notifier cannot find any pages, it doesn't mean no pages are
available...And, if there are no notifiers registered, this condition will
be always true and memory hotplug will show -EBUSY.

This is a bug but not critical.

In most case, a pageblock which will be offlined is MIGRATE_MOVABLE This
"notifier" is called only when the pageblock is _not_ MIGRATE_MOVABLE.
But if not MIGRATE_MOVABLE, it's common case that memory hotplug will
fail.  So, no one notice this bug.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: compaction: fix COMPACTPAGEFAILED counting
Minchan Kim [Tue, 26 Oct 2010 21:21:29 +0000]
mm: compaction: fix COMPACTPAGEFAILED counting

Presently update_nr_listpages() doesn't have a role.  That's because lists
passed is always empty just after calling migrate_pages.  The
migrate_pages cleans up page list which have failed to migrate before
returning by aaa994b3.

 [PATCH] page migration: handle freeing of pages in migrate_pages()

 Do not leave pages on the lists passed to migrate_pages().  Seems that we will
 not need any postprocessing of pages.  This will simplify the handling of
 pages by the callers of migrate_pages().

At that time, we thought we don't need any postprocessing of pages.  But
the situation is changed.  The compaction need to know the number of
failed to migrate for COMPACTPAGEFAILED stat

This patch makes new rule for caller of migrate_pages to call
putback_lru_pages.  So caller need to clean up the lists so it has a
chance to postprocess the pages.  [suggested by Christoph Lameter]

Signed-off-by: Minchan Kim <minchan.kim@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Andi Kleen <andi@firstfloor.org>
Reviewed-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: only build per-node scan_unevictable functions when NUMA is enabled
Thadeu Lima de Souza Cascardo [Tue, 26 Oct 2010 21:21:28 +0000]
mm: only build per-node scan_unevictable functions when NUMA is enabled

Non-NUMA systems do never create these files anyway, since they are only
created by driver subsystem when NUMA is configured.

[akpm@linux-foundation.org: cleanup]
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoinclude/linux/pageblock-flags.h: fix set_pageblock_flags() macro definiton
zeal [Tue, 26 Oct 2010 21:21:27 +0000]
include/linux/pageblock-flags.h: fix set_pageblock_flags() macro definiton

The presently-unused macro was missing one parameter.

Signed-off-by: zeal <zealcook@gmail.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agowriteback: remove nonblocking/encountered_congestion references
Wu Fengguang [Tue, 26 Oct 2010 21:21:26 +0000]
writeback: remove nonblocking/encountered_congestion references

This removes more dead code that was somehow missed by commit 0d99519efef
(writeback: remove unused nonblocking and congestion checks).  There are
no behavior change except for the removal of two entries from one of the
ext4 tracing interface.

The nonblocking checks in ->writepages are no longer used because the
flusher now prefer to block on get_request_wait() than to skip inodes on
IO congestion.  The latter will lead to more seeky IO.

The nonblocking checks in ->writepage are no longer used because it's
redundant with the WB_SYNC_NONE check.

We no long set ->nonblocking in VM page out and page migration, because
a) it's effectively redundant with WB_SYNC_NONE in current code
b) it's old semantic of "Don't get stuck on request queues" is mis-behavior:
   that would skip some dirty inodes on congestion and page out others, which
   is unfair in terms of LRU age.

Inspired by Christoph Hellwig. Thanks!

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: David Howells <dhowells@redhat.com>
Cc: Sage Weil <sage@newdream.net>
Cc: Steve French <sfrench@samba.org>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agooom: fix locking for oom_adj and oom_score_adj
David Rientjes [Tue, 26 Oct 2010 21:21:26 +0000]
oom: fix locking for oom_adj and oom_score_adj

The locking order in oom_adjust_write() and oom_score_adj_write() for
task->alloc_lock and task->sighand->siglock is reversed, and lockdep
notices that irqs could encounter an ABBA scenario.

This fixes the locking order so that we always take task_lock(task) prior
to lock_task_sighand(task).

Signed-off-by: David Rientjes <rientjes@google.com>
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ying Han <yinghan@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agooom: rewrite error handling for oom_adj and oom_score_adj tunables
David Rientjes [Tue, 26 Oct 2010 21:21:25 +0000]
oom: rewrite error handling for oom_adj and oom_score_adj tunables

It's better to use proper error handling in oom_adjust_write() and
oom_score_adj_write() instead of duplicating the locking order on various
exit paths.

Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ying Han <yinghan@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agooom: kill all threads sharing oom killed task's mm
David Rientjes [Tue, 26 Oct 2010 21:21:24 +0000]
oom: kill all threads sharing oom killed task's mm

It's necessary to kill all threads that share an oom killed task's mm if
the goal is to lead to future memory freeing.

This patch reintroduces the code removed in 8c5cd6f3 (oom: oom_kill
doesn't kill vfork parent (or child)) since it is obsoleted.

It's now guaranteed that any task passed to oom_kill_task() does not share
an mm with any thread that is unkillable.  Thus, we're safe to issue a
SIGKILL to any thread sharing the same mm.

This is especially necessary to solve an mm->mmap_sem livelock issue
whereas an oom killed thread must acquire the lock in the exit path while
another thread is holding it in the page allocator while trying to
allocate memory itself (and will preempt the oom killer since a task was
already killed).  Since tasks with pending fatal signals are now granted
access to memory reserves, the thread holding the lock may quickly
allocate and release the lock so that the oom killed task may exit.

This mainly is for threads that are cloned with CLONE_VM but not
CLONE_THREAD, so they are in a different thread group.  Non-NPTL threads
exist in the wild and this change is necessary to prevent the livelock in
such cases.  We care more about preventing the livelock than incurring the
additional tasklist in the oom killer when a task has been killed.
Systems that are sufficiently large to not want the tasklist scan in the
oom killer in the first place already have the option of enabling
/proc/sys/vm/oom_kill_allocating_task, which was designed specifically for
that purpose.

This code had existed in the oom killer for over eight years dating back
to the 2.4 kernel.

[akpm@linux-foundation.org: add nice comment]
Signed-off-by: David Rientjes <rientjes@google.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ying Han <yinghan@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agooom: avoid killing a task if a thread sharing its mm cannot be killed
David Rientjes [Tue, 26 Oct 2010 21:21:23 +0000]
oom: avoid killing a task if a thread sharing its mm cannot be killed

The oom killer's goal is to kill a memory-hogging task so that it may
exit, free its memory, and allow the current context to allocate the
memory that triggered it in the first place.  Thus, killing a task is
pointless if other threads sharing its mm cannot be killed because of its
/proc/pid/oom_adj or /proc/pid/oom_score_adj value.

This patch checks whether any other thread sharing p->mm has an
oom_score_adj of OOM_SCORE_ADJ_MIN.  If so, the thread cannot be killed
and oom_badness(p) returns 0, meaning it's unkillable.

Signed-off-by: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ying Han <yinghan@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agooom: add per-mm oom disable count
Ying Han [Tue, 26 Oct 2010 21:21:23 +0000]
oom: add per-mm oom disable count

It's pointless to kill a task if another thread sharing its mm cannot be
killed to allow future memory freeing.  A subsequent patch will prevent
kills in such cases, but first it's necessary to have a way to flag a task
that shares memory with an OOM_DISABLE task that doesn't incur an
additional tasklist scan, which would make select_bad_process() an O(n^2)
function.

This patch adds an atomic counter to struct mm_struct that follows how
many threads attached to it have an oom_score_adj of OOM_SCORE_ADJ_MIN.
They cannot be killed by the kernel, so their memory cannot be freed in
oom conditions.

This only requires task_lock() on the task that we're operating on, it
does not require mm->mmap_sem since task_lock() pins the mm and the
operation is atomic.

[rientjes@google.com: changelog and sys_unshare() code]
[rientjes@google.com: protect oom_disable_count with task_lock in fork]
[rientjes@google.com: use old_mm for oom_disable_count in exec]
Signed-off-by: Ying Han <yinghan@google.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoDocumentation/filesystems/proc.txt: improve smaps field documentation
Matt Mackall [Tue, 26 Oct 2010 21:21:22 +0000]
Documentation/filesystems/proc.txt: improve smaps field documentation

Signed-off-by: Matt Mackall <mpm@selenic.com>
Cc: Nikanth Karthikesan <knikanth@suse.de>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmcore: it is not experimental any more
WANG Cong [Tue, 26 Oct 2010 21:21:21 +0000]
vmcore: it is not experimental any more

We use vmcore in our production kernel for a long time, it is pretty
stable now.  So I don't think we need to mark it as experimental any more.

Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoum: fix IRQ flag handling naming
Richard Weinberger [Tue, 26 Oct 2010 21:21:21 +0000]
um: fix IRQ flag handling naming

Commit df9ee292 ("Fix IRQ flag handling naming") changed the IRQ flag
handling naming scheme and broke UML:

In file included from arch/um/include/asm/fixmap.h:5,
                 from arch/um/include/shared/um_uaccess.h:10,
                 from arch/um/include/asm/uaccess.h:41,
                 from arch/um/include/asm/thread_info.h:13,
                 from include/linux/thread_info.h:56,
                 from include/linux/preempt.h:9,
                 from include/linux/spinlock.h:50,
                 from include/linux/seqlock.h:29,
                 from include/linux/time.h:8,
                 from include/linux/stat.h:60,
                 from include/linux/module.h:10,
                 from init/main.c:13:
arch/um/include/asm/system.h:11:1: warning: "local_save_flags" redefined

This patch brings the new scheme to UML and makes it work again.

Signed-off-by: Richard Weinberger <richard@nod.at>
Acked-by: David Howells <dhowells@redhat.com>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agopercpu: fix list_head init bug in __percpu_counter_init()
Masanori ITOH [Tue, 26 Oct 2010 21:21:20 +0000]
percpu: fix list_head init bug in __percpu_counter_init()

WARNING: at lib/list_debug.c:26 __list_add+0x3f/0x81()
Hardware name: Express5800/B120a [N8400-085]
list_add corruption. next->prev should be prev (ffffffff81a7ea00), but was dead000000200200. (next=ffff88080b872d58).
Modules linked in: aoe ipt_MASQUERADE iptable_nat nf_nat autofs4 sunrpc bridge 8021q garp stp llc ipv6 cpufreq_ondemand acpi_cpufreq freq_table dm_round_robin dm_multipath kvm_intel kvm uinput lpfc scsi_transport_fc igb ioatdma scsi_tgt i2c_i801 i2c_core dca iTCO_wdt iTCO_vendor_support pcspkr shpchp megaraid_sas [last unloaded: aoe]
Pid: 54, comm: events/3 Tainted: G        W  2.6.34-vanilla1 #1
Call Trace:
[<ffffffff8104bd77>] warn_slowpath_common+0x7c/0x94
[<ffffffff8104bde6>] warn_slowpath_fmt+0x41/0x43
[<ffffffff8120fd2e>] __list_add+0x3f/0x81
[<ffffffff81212a12>] __percpu_counter_init+0x59/0x6b
[<ffffffff810d8499>] bdi_init+0x118/0x17e
[<ffffffff811f2c50>] blk_alloc_queue_node+0x79/0x143
[<ffffffff811f2d2b>] blk_alloc_queue+0x11/0x13
[<ffffffffa02a931d>] aoeblk_gdalloc+0x8e/0x1c9 [aoe]
[<ffffffffa02aa655>] aoecmd_sleepwork+0x25/0xa8 [aoe]
[<ffffffff8106186c>] worker_thread+0x1a9/0x237
[<ffffffffa02aa630>] ? aoecmd_sleepwork+0x0/0xa8 [aoe]
[<ffffffff81065827>] ? autoremove_wake_function+0x0/0x39
[<ffffffff810616c3>] ? worker_thread+0x0/0x237
[<ffffffff810653ad>] kthread+0x7f/0x87
[<ffffffff8100aa24>] kernel_thread_helper+0x4/0x10
[<ffffffff8106532e>] ? kthread+0x0/0x87
[<ffffffff8100aa20>] ? kernel_thread_helper+0x0/0x10

It's because there is no initialization code for a list_head contained in
the struct backing_dev_info under CONFIG_HOTPLUG_CPU, and the bug comes up
when block device drivers calling blk_alloc_queue() are used.  In case of
me, I got them by using aoe.

Signed-off-by: Masanori Itoh <itoumsn@nttdata.co.jp>
Cc: Tejun Heo <tj@kernel.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agokfifo: disable __kfifo_must_check_helper()
Andrew Morton [Tue, 26 Oct 2010 21:21:19 +0000]
kfifo: disable __kfifo_must_check_helper()

This helper is wrong: it coerces signed values into unsigned ones, so code
such as

if (kfifo_alloc(...) < 0) {
error
}

will fail to detect the error.

So let's disable __kfifo_must_check_helper() for 2.6.36.

Cc: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Stefani Seibold <stefani@seibold.net>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agohostfs: fix UML crash: remove f_spare from hostfs
Richard Weinberger [Tue, 26 Oct 2010 21:21:18 +0000]
hostfs: fix UML crash: remove f_spare from hostfs

365b1818 ("add f_flags to struct statfs(64)") resized f_spare within
struct statfs which caused a UML crash.  There is no need to copy f_spare.

Signed-off-by: Richard Weinberger <richard@nod.at>
Reported-by: Toralf Förster <toralf.foerster@gmx.de>
Tested-by: Toralf Förster <toralf.foerster@gmx.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoipmi: proper spinlock initialization
Eric Dumazet [Tue, 26 Oct 2010 21:21:17 +0000]
ipmi: proper spinlock initialization

Unloading ipmi module can trigger following error.  (if
CONFIG_DEBUG_SPINLOCK=y)

[ 9633.779590] BUG: spinlock bad magic on CPU#1, rmmod/7170
[ 9633.779606]  lock: f41f5414, .magic: 00000000, .owner:
<none>/-1, .owner_cpu: 0
[ 9633.779626] Pid: 7170, comm: rmmod Not tainted
2.6.36-rc7-11474-gb71eb1e-dirty #328
[ 9633.779644] Call Trace:
[ 9633.779657]  [<c13921cc>] ? printk+0x18/0x1c
[ 9633.779672]  [<c11a1f33>] spin_bug+0xa3/0xf0
[ 9633.779685]  [<c11a1ffd>] do_raw_spin_lock+0x7d/0x160
[ 9633.779702]  [<c1131537>] ? release_sysfs_dirent+0x47/0xb0
[ 9633.779718]  [<c1131b78>] ? sysfs_addrm_finish+0xa8/0xd0
[ 9633.779734]  [<c1394bac>] _raw_spin_lock_irqsave+0xc/0x20
[ 9633.779752]  [<f99d93da>] cleanup_one_si+0x6a/0x200 [ipmi_si]
[ 9633.779768]  [<c11305b2>] ? sysfs_hash_and_remove+0x72/0x80
[ 9633.779786]  [<f99dcf26>] ipmi_pnp_remove+0xd/0xf [ipmi_si]
[ 9633.779802]  [<c11f622b>] pnp_device_remove+0x1b/0x40

Fix this by initializing spinlocks in a smi_info_alloc() helper function,
right after memory allocation and clearing.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: David Miller <davem@davemloft.net>
Cc: Yinghai Lu <yinghai@kernel.org>
Acked-by: Corey Minyard <cminyard@mvista.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agodrivers/misc/ad525x_dpot.c: fix typo in spi write16 and write24 transfer counts
Michael Hennerich [Tue, 26 Oct 2010 21:21:16 +0000]
drivers/misc/ad525x_dpot.c: fix typo in spi write16 and write24 transfer counts

This is a bug fix.  Some SPI connected devices using 16/24 bit accesses,
previously failed, now work.

This typo slipped in after testing, during some restructuring.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Chris Verges <chrisv@cyberswitching.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoum: remove PAGE_SIZE alignment in linker script causing kernel segfault.
Richard Weinberger [Tue, 26 Oct 2010 21:21:16 +0000]
um: remove PAGE_SIZE alignment in linker script causing kernel segfault.

The linker script cleanup that I did in commit 5d150a97f93 ("um: Clean up
linker script using standard macros.") (2.6.32) accidentally introduced an
ALIGN(PAGE_SIZE) when converting to use INIT_TEXT_SECTION; Richard
Weinberger reported that this causes the kernel to segfault with
CONFIG_STATIC_LINK=y.

I'm not certain why this extra alignment is a problem, but it seems likely
it is because previously

__init_begin = _stext = _text = _sinittext

and with the extra ALIGN(PAGE_SIZE), _sinittext becomes different from the
rest.  So there is likely a bug here where something is assuming that
_sinittext is the same as one of those other symbols.  But reverting the
accidental change fixes the regression, so it seems worth committing that
now.

Signed-off-by: Tim Abbott <tabbott@ksplice.com>
Reported-by: Richard Weinberger <richard@nod.at>
Cc: Jeff Dike <jdike@addtoit.com>
Tested by: Antoine Martin <antoine@nagafix.co.uk>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agosgi-xp: incoming XPC channel messages can come in after the channel's partition struc...
Robin Holt [Tue, 26 Oct 2010 21:21:15 +0000]
sgi-xp: incoming XPC channel messages can come in after the channel's partition structures have been torn down

Under some workloads, some channel messages have been observed being
delayed on the sending side past the point where the receiving side has
been able to tear down its partition structures.

This condition is already detected in xpc_handle_activate_IRQ_uv(), but
that information is not given to xpc_handle_activate_mq_msg_uv().  As a
result, xpc_handle_activate_mq_msg_uv() assumes the structures still exist
and references them, causing a NULL-pointer deref.

Signed-off-by: Robin Holt <holt@sgi.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoum: fix global timer issue when using CONFIG_NO_HZ
Richard Weinberger [Tue, 26 Oct 2010 21:21:13 +0000]
um: fix global timer issue when using CONFIG_NO_HZ

This fixes a issue which was introduced by fe2cc53e ("uml: track and make
up lost ticks").

timeval_to_ns() returns long long and not int.  Due to that UML's timer
did not work properlt and caused timer freezes.

Signed-off-by: Richard Weinberger <richard@nod.at>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm, page-allocator: do not check the state of a non-existant buddy during free
Mel Gorman [Tue, 26 Oct 2010 21:21:11 +0000]
mm, page-allocator: do not check the state of a non-existant buddy during free

There is a bug in commit 6dda9d55 ("page allocator: reduce fragmentation
in buddy allocator by adding buddies that are merging to the tail of the
free lists") that means a buddy at order MAX_ORDER is checked for merging.
 A page of this order never exists so at times, an effectively random
piece of memory is being checked.

Alan Curry has reported that this is causing memory corruption in
userspace data on a PPC32 platform (http://lkml.org/lkml/2010/10/9/32).
It is not clear why this is happening.  It could be a cache coherency
problem where pages mapped in both user and kernel space are getting
different cache lines due to the bad read from kernel space
(http://lkml.org/lkml/2010/10/13/179).  It could also be that there are
some special registers being io-remapped at the end of the memmap array
and that a read has special meaning on them.  Compiler bugs have been
ruled out because the assembly before and after the patch looks relatively
harmless.

This patch fixes the problem by ensuring we are not reading a possibly
invalid location of memory.  It's not clear why the read causes corruption
but one way or the other it is a buggy read.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Corrado Zoccolo <czoccolo@gmail.com>
Reported-by: Alan Curry <pacman@kosh.dhis.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agotypes.h: move misplaced comment
Andrew Morton [Tue, 26 Oct 2010 21:21:10 +0000]
types.h: move misplaced comment

This comment landed in the wrong place.

Cc: Andi Kleen <andi@firstfloor.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: David Miller <davem@davemloft.net>
Cc: Eric Paris <eparis@redhat.com>
Cc: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: fix return value of scan_lru_pages in memory unplug
KAMEZAWA Hiroyuki [Tue, 26 Oct 2010 21:21:10 +0000]
mm: fix return value of scan_lru_pages in memory unplug

scan_lru_pages returns pfn. So, it's type should be "unsigned long"
not "int".

Note: I guess this has been work until now because memory hotplug tester's
      machine has not very big memory....
      physical address < 32bit << PAGE_SHIFT.

Reported-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoMerge git://git.infradead.org/battery-2.6
Linus Torvalds [Tue, 26 Oct 2010 17:14:23 +0000]
Merge git://git.infradead.org/battery-2.6

* git://git.infradead.org/battery-2.6:
  power_supply: Makefile cleanup
  bq27x00_battery: Add missing kfree(di->bus) in bq27x00_battery_remove()
  power_supply: Introduce maximum current property
  power_supply: Add types for USB chargers
  ds2782_battery: Fix units
  power_supply: Add driver for TWL4030/TPS65950 BCI charger
  bq20z75: Add support for more power supply properties
  wm831x_power: Add missing kfree(wm831x_power) in wm831x_power_remove()
  jz4740-battery: Add missing kfree(jz_battery) in jz_battery_remove()
  ds2760_battery: Add missing kfree(di) in ds2760_battery_remove()
  olpc_battery: Fix endian neutral breakage for s16 values
  ds2760_battery: Fix W1 and W1_SLAVE_DS2760 dependency
  pcf50633-charger: Add missing sysfs_remove_group()
  power_supply: Add driver for TI BQ20Z75 gas gauge IC
  wm831x_power: Remove duplicate chg mask
  omap: rx51: Add support for USB chargers
  power_supply: Add isp1704 charger detection driver

8 years agoMerge branch 'linux_next' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Tue, 26 Oct 2010 17:13:48 +0000]
Merge branch 'linux_next' of git://git./linux/kernel/git/mchehab/i7core

* 'linux_next' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/i7core: (34 commits)
  i7core_edac: return -ENODEV when devices were already probed
  i7core_edac: properly terminate pci_dev_table
  i7core_edac: Avoid PCI refcount to reach zero on successive load/reload
  i7core_edac: Fix refcount error at PCI devices
  i7core_edac: it is safe to i7core_unregister_mci() when mci=NULL
  i7core_edac: Fix an oops at i7core probe
  i7core_edac: Remove unused member channels in i7core_pvt
  i7core_edac: Remove unused arg csrow from get_dimm_config
  i7core_edac: Reduce args of i7core_register_mci
  i7core_edac: Introduce i7core_unregister_mci
  i7core_edac: Use saved pointers
  i7core_edac: Check probe counter in i7core_remove
  i7core_edac: Call pci_dev_put() when alloc_i7core_dev()  failed
  i7core_edac: Fix error path of i7core_register_mci
  i7core_edac: Fix order of lines in i7core_register_mci
  i7core_edac: Always do get/put for all devices
  i7core_edac: Introduce i7core_pci_ctl_create/release
  i7core_edac: Introduce free_i7core_dev
  i7core_edac: Introduce alloc_i7core_dev
  i7core_edac: Reduce args of i7core_get_onedevice
  ...

8 years agoMerge branch 'hwpoison' of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux...
Linus Torvalds [Tue, 26 Oct 2010 17:13:10 +0000]
Merge branch 'hwpoison' of git://git./linux/kernel/git/ak/linux-mce-2.6

* 'hwpoison' of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-mce-2.6: (22 commits)
  Add _addr_lsb field to ia64 siginfo
  Fix migration.c compilation on s390
  HWPOISON: Remove retry loop for try_to_unmap
  HWPOISON: Turn addr_valid from bitfield into char
  HWPOISON: Disable DEBUG by default
  HWPOISON: Convert pr_debugs to pr_info
  HWPOISON: Improve comments in memory-failure.c
  x86: HWPOISON: Report correct address granuality for huge hwpoison faults
  Encode huge page size for VM_FAULT_HWPOISON errors
  Fix build error with !CONFIG_MIGRATION
  hugepage: move is_hugepage_on_freelist inside ifdef to avoid warning
  Clean up __page_set_anon_rmap
  HWPOISON, hugetlb: fix unpoison for hugepage
  HWPOISON, hugetlb: soft offlining for hugepage
  HWPOSION, hugetlb: recover from free hugepage error when !MF_COUNT_INCREASED
  hugetlb: move refcounting in hugepage allocation inside hugetlb_lock
  HWPOISON, hugetlb: add free check to dequeue_hwpoison_huge_page()
  hugetlb: hugepage migration core
  hugetlb: redefine hugepage copy functions
  hugetlb: add allocate function for hugepage migration
  ...

8 years agoMerge branch 'for_linus' of git://github.com/at91linux/linux-2.6-at91
Linus Torvalds [Tue, 26 Oct 2010 17:03:40 +0000]
Merge branch 'for_linus' of git://github.com/at91linux/linux-2.6-at91

* 'for_linus' of git://github.com/at91linux/linux-2.6-at91:
  AT91: rtc: enable built-in RTC in Kconfig for at91sam9g45 family
  at91/atmel-mci: inclusion of sd/mmc driver in at91sam9g45 chip and board
  AT91: pm: make sure that r0 is 0 when dealing with cache operations
  AT91: pm: use plain cpu_do_idle() for "wait for interrupt"
  AT91: reset: extend alternate reset procedure to several chips
  AT91: reset routine cleanup, remove not needed icache flush
  AT91: trivial: align comment of at91sam9g20_reset with one more tab
  AT91: Fix AT91SAM9G20 reset as per the errata in the data sheet
  AT91: add board support for Pcontrol_G20

8 years agoMerge branch 'for-linus' of git://gitorious.org/linux-omap-dss2/linux
Linus Torvalds [Tue, 26 Oct 2010 17:02:39 +0000]
Merge branch 'for-linus' of git://gitorious.org/linux-omap-dss2/linux

* 'for-linus' of git://gitorious.org/linux-omap-dss2/linux:
  OMAP: DSS2: don't power off a panel twice
  OMAP: DSS2: OMAPFB: Allow usage of def_vrfb only for omap2,3
  OMAP: DSS2: OMAPFB: make VRFB depends on OMAP2,3
  OMAP: DSS2: OMAPFB: Allow FB_OMAP2 to build without VRFB
  arm/omap: simplify conditional
  OMAP: DSS2: DSI: Remove extra iounmap in error path
  OMAP: DSS2: Use dss_features framework on DSS2 code
  OMAP: DSS2: Introduce dss_features files
  video/omap: remove mux.h include
  ARM: omap/fb: move get_fbmem_region() to .init.text
  ARM: omap/fb: move omapfb_reserve_sram to .init.text
  ARM: omap/fb: move omap_init_fb to .init.text
  OMAP: DSS2: OMAPFB: swap front and back porches for both hsync and vsync
  OMAP: DSS2: make filter coefficient tables human readable
  OMAP: DSS2: Add SPI dependency to Kconfig of ACX565AKM panel

8 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq
Linus Torvalds [Tue, 26 Oct 2010 17:00:04 +0000]
Merge branch 'next' of git://git./linux/kernel/git/davej/cpufreq

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq:
  [CPUFREQ]: x86, cpufreq: Mark longrun_get_policy with __cpuinit.
  [CPUFREQ] add sampling_down_factor tunable to improve ondemand performance
  [CPUFREQ] arch/x86/kernel/cpu/cpufreq: Fix unsigned return type
  [CPUFREQ] drivers/cpufreq: Adjust confusing if indentation

8 years agoMerge branch 'for-2.6.37' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Tue, 26 Oct 2010 16:55:25 +0000]
Merge branch 'for-2.6.37' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.37' of git://linux-nfs.org/~bfields/linux: (99 commits)
  svcrpc: svc_tcp_sendto XPT_DEAD check is redundant
  svcrpc: no need for XPT_DEAD check in svc_xprt_enqueue
  svcrpc: assume svc_delete_xprt() called only once
  svcrpc: never clear XPT_BUSY on dead xprt
  nfsd4: fix connection allocation in sequence()
  nfsd4: only require krb5 principal for NFSv4.0 callbacks
  nfsd4: move minorversion to client
  nfsd4: delay session removal till free_client
  nfsd4: separate callback change and callback probe
  nfsd4: callback program number is per-session
  nfsd4: track backchannel connections
  nfsd4: confirm only on succesful create_session
  nfsd4: make backchannel sequence number per-session
  nfsd4: use client pointer to backchannel session
  nfsd4: move callback setup into session init code
  nfsd4: don't cache seq_misordered replies
  SUNRPC: Properly initialize sock_xprt.srcaddr in all cases
  SUNRPC: Use conventional switch statement when reclassifying sockets
  sunrpc/xprtrdma: clean up workqueue usage
  sunrpc: Turn list_for_each-s into the ..._entry-s
  ...

Fix up trivial conflicts (two different deprecation notices added in
separate branches) in Documentation/feature-removal-schedule.txt

8 years agoMerge branch 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6
Linus Torvalds [Tue, 26 Oct 2010 16:52:09 +0000]
Merge branch 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
  net/sunrpc: Use static const char arrays
  nfs4: fix channel attribute sanity-checks
  NFSv4.1: Use more sensible names for 'initialize_mountpoint'
  NFSv4.1: pnfs: filelayout: add driver's LAYOUTGET and GETDEVICEINFO infrastructure
  NFSv4.1: pnfs: add LAYOUTGET and GETDEVICEINFO infrastructure
  NFS: client needs to maintain list of inodes with active layouts
  NFS: create and destroy inode's layout cache
  NFSv4.1: pnfs: filelayout: introduce minimal file layout driver
  NFSv4.1: pnfs: full mount/umount infrastructure
  NFS: set layout driver
  NFS: ask for layouttypes during v4 fsinfo call
  NFS: change stateid to be a union
  NFSv4.1: pnfsd, pnfs: protocol level pnfs constants
  SUNRPC: define xdr_decode_opaque_fixed
  NFSD: remove duplicate NFS4_STATEID_SIZE

8 years agoAT91: rtc: enable built-in RTC in Kconfig for at91sam9g45 family
Nicolas Ferre [Fri, 22 Oct 2010 17:12:52 +0000]
AT91: rtc: enable built-in RTC in Kconfig for at91sam9g45 family

Enable built-in RTC IP in Kconfig and modify comments and help messages.
RTT as RTC is still available but should not be selected in common case.

Reported-by: Yegor Yefremov <yegor_sub1@visionsystems.de>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoat91/atmel-mci: inclusion of sd/mmc driver in at91sam9g45 chip and board
Nicolas Ferre [Fri, 22 Oct 2010 16:27:48 +0000]
at91/atmel-mci: inclusion of sd/mmc driver in at91sam9g45 chip and board

This adds the support of atmel-mci sd/mmc driver in at91sam9g45 devices and
board files. This also configures the DMA controller slave interface for
at_hdmac dmaengine driver.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: pm: make sure that r0 is 0 when dealing with cache operations
Nicolas Ferre [Fri, 22 Oct 2010 16:55:39 +0000]
AT91: pm: make sure that r0 is 0 when dealing with cache operations

When using CP15 cache operations (c7), we make sure that Rd (r0)
is actually 0 as ARM 926 TRM is saying.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: pm: use plain cpu_do_idle() for "wait for interrupt"
Nicolas Ferre [Fri, 22 Oct 2010 15:53:39 +0000]
AT91: pm: use plain cpu_do_idle() for "wait for interrupt"

For power management at91_pm_enter() routine, use the cpu_do_idle() for a
rock solid "wait for interrupt" implementation.
For AT91SAM9 ARM 926 based chips, we can exceed the cache line length as
we can access RAM even while in self-refresh mode.
We keep plain access to CP15 for at91rm9200 as this feature is not
available: instructions have to be in a single cache line.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: reset: extend alternate reset procedure to several chips
Nicolas Ferre [Thu, 14 Oct 2010 17:14:00 +0000]
AT91: reset: extend alternate reset procedure to several chips

Several at91sam9 chips need the alternate reset procedure to be sure to halt
SDRAM smoothly before resetting the chip.
This is an extension of previous patch "Fix AT91SAM9G20 reset" to all chips
affected.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: reset routine cleanup, remove not needed icache flush
Nicolas Ferre [Thu, 14 Oct 2010 15:19:11 +0000]
AT91: reset routine cleanup, remove not needed icache flush

Generalize assembler reset routine to allow use on several at91sam9 chips.
This patch replace double definitions of SDRAM controller registers and RSTC
registers with use of classical header files.

For this rework, we remove the not needed icache flush as it is already
done in the calling function: arm_machine_restart().

Rename at91sam9g20_reset.S to generalize to several chips.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: trivial: align comment of at91sam9g20_reset with one more tab
Nicolas Ferre [Thu, 14 Oct 2010 14:51:36 +0000]
AT91: trivial: align comment of at91sam9g20_reset with one more tab

Preparing next patch with longer names

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: Fix AT91SAM9G20 reset as per the errata in the data sheet
Peter Horton [Fri, 28 May 2010 15:37:26 +0000]
AT91: Fix AT91SAM9G20 reset as per the errata in the data sheet

If the SDRAM is not cleanly shutdown before reset it can be left driving
the bus, which then stops the bootloader booting from NAND.

Signed-off-by: Peter Horton <phorton@bitbox.co.uk>
[nicolas.ferre@atmel.com: change file header line order]
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agoAT91: add board support for Pcontrol_G20
Peter Gsellmann [Wed, 13 Oct 2010 14:18:51 +0000]
AT91: add board support for Pcontrol_G20

Board is a carrier board for Stamp9G20, with additional peripherals
for a building automation system

Signed-off-by: Peter Gsellmann <pgsellmann@portner-elektronik.at>
[nicolas.ferre@atmel.com: remove machine_desc.io_pg_offst and .phys_io]
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>

8 years agonet/sunrpc: Use static const char arrays
Joe Perches [Mon, 13 Sep 2010 19:48:01 +0000]
net/sunrpc: Use static const char arrays

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

8 years agonfs4: fix channel attribute sanity-checks
J. Bruce Fields [Sat, 2 Oct 2010 19:19:01 +0000]
nfs4: fix channel attribute sanity-checks

The sanity checks here are incorrect; in the worst case they allow
values that crash the client.

They're also over-reliant on the preprocessor.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

8 years agoMerge branch 'for-next' of git://android.git.kernel.org/kernel/tegra
Linus Torvalds [Tue, 26 Oct 2010 01:42:06 +0000]
Merge branch 'for-next' of git://android.git./kernel/tegra

* 'for-next' of git://android.git.kernel.org/kernel/tegra:
  spi: tegra: fix error setting on timeout
  spi: add spi_tegra driver
  tegra: harmony: enable PCI Express
  tegra: add PCI Express support
  tegra: add PCI Express clocks
  [ARM] tegra: Add APB DMA support
  [ARM] tegra: Add cpufreq support
  [ARM] tegra: common: Update common clock init table
  [ARM] tegra: clock: Add dvfs support, bug fixes, and cleanups
  [ARM] tegra: Add support for reading fuses
  [ARM] tegra: gpio: Add suspend and wake support
  [ARM] tegra: pinmux: add safe values, move tegra2, add suspend
  [ARM] tegra: add suspend and mirror irqs to legacy controller
  [ARM] tegra: Add legacy irq support
  [ARM] tegra: update iomap

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/vapier...
Linus Torvalds [Tue, 26 Oct 2010 01:41:32 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/vapier/blackfin

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/vapier/blackfin:
  Blackfin: fix inverted anomaly 05000481 logic
  Blackfin: drop unused irq_panic()/DEBUG_ICACHE_CHECK
  Blackfin: ppi/spi/twi headers: add missing __BFP undef
  Blackfin: update defconfigs
  Blackfin: bfin_twi.h: start a common TWI header
  netdev: bfin_mac: push settings to platform resources

8 years agospi: tegra: fix error setting on timeout
Erik Gilling [Thu, 9 Sep 2010 01:16:45 +0000]
spi: tegra: fix error setting on timeout

avoids derefencing an uninitialized pointer

Change-Id: Icf528441ae481e9f6f5ddc0be32c7c217fa49701
Signed-off-by: Erik Gilling <konkers@android.com>

8 years agoMerge branch 'next' of git://git.monstr.eu/linux-2.6-microblaze
Linus Torvalds [Mon, 25 Oct 2010 23:53:11 +0000]
Merge branch 'next' of git://git.monstr.eu/linux-2.6-microblaze

* 'next' of git://git.monstr.eu/linux-2.6-microblaze: (42 commits)
  microblaze: Fix build with make 3.82
  fbdev/xilinxfb: Microblaze driver support
  microblaze: Support C optimized lib functions for little-endian
  microblaze: Separate library optimized functions
  microblaze: Support timer on AXI lite
  microblaze: Add support for little-endian Microblaze
  microblaze: KGDB little endian support
  microblaze: Add PVR for endians plus detection
  net: emaclite: Add support for little-endian platforms
  microblaze: trivial: Add comment for AXI pvr
  microblaze: pci-common cleanup
  microblaze: Support early console on uart16550
  microblaze: Do not compile early console support for uartlite if is disabled
  microblaze: Setup early console dynamically
  microblaze: Rename all uartlite early printk functions
  microblaze: remove early printk uarlite console dependency from header
  microblaze: Remove additional compatible properties
  microblaze: Remove hardcoded asm instraction for PVR loading
  microblaze: Use static const char * const where possible
  microblaze: Define VMALLOC_START/END
  ...

8 years agoMerge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groec...
Linus Torvalds [Mon, 25 Oct 2010 23:25:31 +0000]
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/staging

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging: (24 commits)
  hwmon: lis3: Release resources in case of failure
  hwmon: lis3: Short explanations of platform data fields
  hwmon: lis3: Enhance lis3 selftest with IRQ line test
  hwmon: lis3: use block read to access data registers
  hwmon: lis3: Adjust fuzziness for 8 bit device
  hwmon: lis3: New parameters to platform data
  hwmon: lis3: restore axis enabled bits
  hwmon: lis3: Power on corrections
  hwmon: lis3: Update coordinates at polled device open
  hwmon: lis3: Cleanup interrupt handling
  hwmon: lis3: regulator control
  hwmon: lis3: pm_runtime support
  Kirkwood: add fan support for Network Space Max v2
  hwmon: add generic GPIO fan driver
  hwmon: (coretemp) fix reading of microcode revision (v2)
  hwmon: ({core, pkg, via-cpu}temp) remove unnecessary CONFIG_HOTPLUG_CPU ifdefs
  hwmon: (pkgtemp) align driver initialization style with coretemp
  hwmon: LTC4261 Hardware monitoring driver
  hwmon: (lis3) add axes module parameter for custom axis-mapping
  hwmon: (hp_accel) Add HP Mini 510x family support
  ...

8 years agoMN10300: Fix the PERCPU() alignment to allow for workqueues
David Howells [Mon, 25 Oct 2010 22:41:11 +0000]
MN10300: Fix the PERCPU() alignment to allow for workqueues

In the MN10300 arch, we occasionally see an assertion being tripped in
alloc_cwqs() at the following line:

        /* just in case, make sure it's actually aligned */
  --->  BUG_ON(!IS_ALIGNED(wq->cpu_wq.v, align));
        return wq->cpu_wq.v ? 0 : -ENOMEM;

The values are:

        wa->cpu_wq.v => 0x902776e0
        align => 0x100

and align is calculated by the following:

        const size_t align = max_t(size_t, 1 << WORK_STRUCT_FLAG_BITS,
                                   __alignof__(unsigned long long));

This is because the pointer in question (wq->cpu_wq.v) loses some of its
lower bits to control flags, and so the object it points to must be
sufficiently aligned to avoid the need to use those bits for pointing to
things.

Currently, 4 control bits and 4 colour bits are used in normal
circumstances, plus a debugging bit if debugging is set.  This requires
the cpu_workqueue_struct struct to be at least 256 bytes aligned (or 512
bytes aligned with debugging).

PERCPU() alignment on MN13000, however, is only 32 bytes as set in
vmlinux.lds.S.  So we set this to PAGE_SIZE (4096) to match most other
arches and stick a comment in alloc_cwqs() for anyone else who triggers
the assertion.

Reported-by: Akira Takeuchi <takeuchi.akr@jp.panasonic.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Mark Salter <msalter@redhat.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agosvcrpc: svc_tcp_sendto XPT_DEAD check is redundant
J. Bruce Fields [Sat, 23 Oct 2010 15:55:53 +0000]
svcrpc: svc_tcp_sendto XPT_DEAD check is redundant

The only caller (svc_send) has already checked XPT_DEAD.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>

8 years agosvcrpc: no need for XPT_DEAD check in svc_xprt_enqueue
J. Bruce Fields [Sat, 23 Oct 2010 15:23:46 +0000]
svcrpc: no need for XPT_DEAD check in svc_xprt_enqueue

If any xprt marked DEAD is also left BUSY for the rest of its life, then
the XPT_DEAD check here is superfluous--we'll get the same result from
the XPT_BUSY check just after.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>

8 years agosvcrpc: assume svc_delete_xprt() called only once
J. Bruce Fields [Sat, 23 Oct 2010 15:16:10 +0000]
svcrpc: assume svc_delete_xprt() called only once

As long as DEAD exports are left BUSY, and svc_delete_xprt is called
only with BUSY held, then svc_delete_xprt() will never be called on an
xprt that is already DEAD.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>

8 years agosvcrpc: never clear XPT_BUSY on dead xprt
J. Bruce Fields [Sat, 23 Oct 2010 15:04:12 +0000]
svcrpc: never clear XPT_BUSY on dead xprt

Once an xprt has been deleted, there's no reason to allow it to be
enqueued--at worst, that might cause the xprt to be re-added to some
global list, resulting in later corruption.

Also, note this leaves us with no need for the reference-count
manipulation here.

Reviewed-by: NeilBrown <neilb@suse.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>

8 years agohwmon: lis3: Release resources in case of failure
Samu Onkalo [Fri, 22 Oct 2010 11:57:34 +0000]
hwmon: lis3: Release resources in case of failure

If lis3lv02d_init_device fails, HW resources were not released
properly. In case of failure call release_resources if available.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agoBlackfin: fix inverted anomaly 05000481 logic
Mike Frysinger [Mon, 25 Oct 2010 08:49:49 +0000]
Blackfin: fix inverted anomaly 05000481 logic

No one uses these MMRs so we didn't notice when the anomaly handling
logic was inverted.

Reported-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>

8 years agoBlackfin: drop unused irq_panic()/DEBUG_ICACHE_CHECK
Mike Frysinger [Mon, 25 Oct 2010 21:22:49 +0000]
Blackfin: drop unused irq_panic()/DEBUG_ICACHE_CHECK

This code was useful during early port development when our icache code
wasn't solid, but that ship has sailed long ago, and no code calls this
function anymore (irq_panic).  So punt it.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>

8 years agohwmon: lis3: Short explanations of platform data fields
Samu Onkalo [Sat, 23 Oct 2010 13:39:44 +0000]
hwmon: lis3: Short explanations of platform data fields

Short documentation at kernel doc format.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Jonathan Cameron <jic23@cam.ac.uk>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: Enhance lis3 selftest with IRQ line test
Samu Onkalo [Fri, 22 Oct 2010 11:57:32 +0000]
hwmon: lis3: Enhance lis3 selftest with IRQ line test

Configure chip to data ready mode in selftest and count received
interrupts to see that interrupt line(s) are working.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: use block read to access data registers
Samu Onkalo [Fri, 22 Oct 2010 11:57:31 +0000]
hwmon: lis3: use block read to access data registers

Add optional blockread function to interface driver. If available
the chip driver uses it for data register access. For 12 bit device
it reads 6 bytes to get 3*16bit data. For 8 bit device it reads out
5 bytes since every second byte is dummy.
This optimizes bus usage and reduces number of operations and
interrupts needed for one data update.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Jonathan Cameron <jic23@cam.ac.uk>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: Adjust fuzziness for 8 bit device
Samu Onkalo [Fri, 22 Oct 2010 11:57:30 +0000]
hwmon: lis3: Adjust fuzziness for 8 bit device

Default fuziness is set smaller for 8 device.
In 12 bit device LSB is quite close to 1 mg
(mg = 1 / 1000 of earth gravity).
In 8bit device LSB is about 18 mg.
Set fuziness to 1 for 8 bit device.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: New parameters to platform data
Samu Onkalo [Fri, 22 Oct 2010 11:57:29 +0000]
hwmon: lis3: New parameters to platform data

Added default output data rate setting to platform data.
If default rate is 0, reset default value is used.
Added control for duration via platform data.
Added possibility to configure interrupts to trig on
both rising and falling edge. The lis3 WU unit can be
configured quite many ways and with some configurations it
is quite handy to get coordinate refresh when some
event trigs and when it reason goes away.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Jonathan Cameron <jic23@cam.ac.uk>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: restore axis enabled bits
Samu Onkalo [Fri, 22 Oct 2010 11:57:28 +0000]
hwmon: lis3: restore axis enabled bits

All axis enable bits are set to 0 at module remove.
Restore reset default value at init.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: Power on corrections
Samu Onkalo [Fri, 22 Oct 2010 11:57:27 +0000]
hwmon: lis3: Power on corrections

Sometimes lis3 chip seems to fail to setup factory tuning at boot up.
This probably happens if there is some odd power ramp down ramp up sequence
for example in device restart. Set boot bit in control2 register to
trig boot sequence manually and wait until it is finished.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Jonathan Cameron <jic23@cam.ac.uk>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: Update coordinates at polled device open
Samu Onkalo [Fri, 22 Oct 2010 11:57:26 +0000]
hwmon: lis3: Update coordinates at polled device open

Call input device poll function at device open to refresh coordinates
immediately. This is needed for the case where poll interval is set to
zero and coordinate updates happens purely under interrupt control.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: Cleanup interrupt handling
Samu Onkalo [Fri, 22 Oct 2010 11:57:25 +0000]
hwmon: lis3: Cleanup interrupt handling

Irqcfg moved to chip data instead of platform data.
This simplifies access in interrupt handler little bit.

Input device open and close functions set status for
interrupt threaded handler once.

Unnecessary check for interrupt source removed since
it is enough that active interrupt line indicates that
there was an interrupt.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: regulator control
Samu Onkalo [Fri, 22 Oct 2010 11:57:24 +0000]
hwmon: lis3: regulator control

Based on pm_runtime control, turn lis3 regulators on and off.
Perform context save and restore on transitions.

Feature is optional and must be enabled in platform data.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Jonathan Cameron <jic23@cam.ac.uk>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: lis3: pm_runtime support
Samu Onkalo [Fri, 22 Oct 2010 11:57:23 +0000]
hwmon: lis3: pm_runtime support

Add pm_runtime support to lis3 core driver.
Add pm_runtime support to lis3 i2c driver.

spi and hp_accel drivers are not yet supported. Old always
on functionality remains for those.

For sysfs there is 5 second delay before turning off the
chip to avoid long ramp up delay.

Signed-off-by: Samu Onkalo <samu.p.onkalo@nokia.com>
Acked-by: Jonathan Cameron <jic23@cam.ac.uk>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agoKirkwood: add fan support for Network Space Max v2
Simon Guinot [Fri, 22 Oct 2010 09:29:18 +0000]
Kirkwood: add fan support for Network Space Max v2

Signed-off-by: Simon Guinot <sguinot@lacie.com>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: add generic GPIO fan driver
Simon Guinot [Thu, 21 Oct 2010 22:44:19 +0000]
hwmon: add generic GPIO fan driver

This patch adds hwmon support for fans connected to GPIO lines.

Platform specific information such as GPIO pinout and speed conversion array
(rpm from/to GPIO value) are passed to the driver via platform_data.

Signed-off-by: Simon Guinot <sguinot@lacie.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: (coretemp) fix reading of microcode revision (v2)
Jan Beulich [Fri, 8 Oct 2010 08:59:38 +0000]
hwmon: (coretemp) fix reading of microcode revision (v2)

According to the documentation, simply reading the respective MSR
isn't sufficient: It should be written with zeros, cpuid(1) be
executed, and then read (see arch/x86/kernel/cpu/intel.c for an
example).

v2: Fail probe when microcode revision cannot be determined, but is
needed to check for proper operation.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Chen Gong <gong.chen@linux.intel.com>
Cc: Jean Delvare <khali@linux-fr.org>
Acked-by: Fenghua Yu <fenghua.yu@intel.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: ({core, pkg, via-cpu}temp) remove unnecessary CONFIG_HOTPLUG_CPU ifdefs
Chen Gong [Sat, 9 Oct 2010 02:01:48 +0000]
hwmon: ({core, pkg, via-cpu}temp) remove unnecessary CONFIG_HOTPLUG_CPU ifdefs

CONFIG_HOTPLUG_CPU is used too much in some drivers.
This patch clean them up.

Signed-off-by: Chen Gong <gong.chen@linux.intel.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: (pkgtemp) align driver initialization style with coretemp
Chen Gong [Fri, 8 Oct 2010 05:53:35 +0000]
hwmon: (pkgtemp) align driver initialization style with coretemp

pkgtemp is derived from coretemp, so some reasonable
logics should be applied onto pkgtemp, too. Such as
the init logic here.

Signed-off-by: Chen Gong <gong.chen@linux.intel.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: LTC4261 Hardware monitoring driver
Guenter Roeck [Fri, 25 Jun 2010 18:59:54 +0000]
hwmon: LTC4261 Hardware monitoring driver

This driver adds support for Linear Technology LTC4261 I2C Negative
Voltage Hot Swap Controller.

Reviewed-by: Ira W. Snyder <iws@ovro.caltech.edu>
Reviewed-by: Tom Grennan <tom.grennan@ericsson.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: (lis3) add axes module parameter for custom axis-mapping
Takashi Iwai [Fri, 1 Oct 2010 21:14:25 +0000]
hwmon: (lis3) add axes module parameter for custom axis-mapping

The axis-mapping of lis3dev device on many (rather most) HP machines
doesn't follow the standard.  When each new model appears, users need to
adjust again.  Testing this requires the rebuild of kernel, thus it's not
trivial for end-users.

This patch adds a module parameter "axes" to allow a custom axis-mapping
without patching and recompiling the kernel driver.  User can pass the
parameter such as axes=3,2,1.  Also it can be changed via sysfs.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Acked-by: Eric Piel <eric.piel@tremplin-utc.net>
Cc: Jean Delvare <khali@linux-fr.org>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: (hp_accel) Add HP Mini 510x family support
Masanari Iida [Fri, 27 Aug 2010 00:21:43 +0000]
hwmon: (hp_accel) Add HP Mini 510x family support

This patch is an enhanced version of Takashi Iwai's
[PATCH] hp_accel: Add quirks for HP ProBook 532x and HP Mini 5102
My HP Mini 5101 works fine with this patch.
Confirmed with Tux Racer.

Signed-off by: Masanari Iida <standby24x7@gmail.com>
Acked-by: Takashi Iwai <tiwai@suse.de>
Cc: Eric Piel <eric.piel@tremplin-utc.net>

Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agohwmon: (lis3) Add support for new LIS3DC / HP3DC chip
Takashi Iwai [Thu, 23 Sep 2010 17:01:39 +0000]
hwmon: (lis3) Add support for new LIS3DC / HP3DC chip

A new version of LIS3 chip has slight incompatibilities from former
versions.  This patch adds the minimal support for it.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agox86/hwmon: remove inclusion of unnecessary headers from {core, pkg, via-cpu}temp.c
Jan Beulich [Mon, 13 Sep 2010 10:32:08 +0000]
x86/hwmon: remove inclusion of unnecessary headers from {core, pkg, via-cpu}temp.c

These likely originate from these drivers being clones of one another
and/or other drivers which actually needed these includes.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agox86/hwmon: (coretemp) cosmetic cleanup
Jan Beulich [Mon, 13 Sep 2010 10:28:35 +0000]
x86/hwmon: (coretemp) cosmetic cleanup

"break" after "return" is at best bogus (good compilers even warn about
the "break" being unreachable).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agox86/hwmon: {core, pkg, via}cpu_temp_device_remove() can all be __cpuinit
Jan Beulich [Fri, 24 Sep 2010 05:31:10 +0000]
x86/hwmon: {core, pkg, via}cpu_temp_device_remove() can all be __cpuinit

... as they're being called only from a __cpuinit function.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

8 years agoMerge branch 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6
Linus Torvalds [Mon, 25 Oct 2010 20:48:29 +0000]
Merge branch 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6: (67 commits)
  SUNRPC: Cleanup duplicate assignment in rpcauth_refreshcred
  nfs: fix unchecked value
  Ask for time_delta during fsinfo probe
  Revalidate caches on lock
  SUNRPC: After calling xprt_release(), we must restart from call_reserve
  NFSv4: Fix up the 'dircount' hint in encode_readdir
  NFSv4: Clean up nfs4_decode_dirent
  NFSv4: nfs4_decode_dirent must clear entry->fattr->valid
  NFSv4: Fix a regression in decode_getfattr
  NFSv4: Fix up decode_attr_filehandle() to handle the case of empty fh pointer
  NFS: Ensure we check all allocation return values in new readdir code
  NFS: Readdir plus in v4
  NFS: introduce generic decode_getattr function
  NFS: check xdr_decode for errors
  NFS: nfs_readdir_filler catch all errors
  NFS: readdir with vmapped pages
  NFS: remove page size checking code
  NFS: decode_dirent should use an xdr_stream
  SUNRPC: Add a helper function xdr_inline_peek
  NFS: remove readdir plus limit
  ...

8 years agoMerge branch 'omap-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind...
Linus Torvalds [Mon, 25 Oct 2010 20:46:56 +0000]
Merge branch 'omap-for-linus' of git://git./linux/kernel/git/tmlind/linux-omap-2.6

* 'omap-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6: (163 commits)
  omap: complete removal of machine_desc.io_pg_offst and .phys_io
  omap: UART: fix wakeup registers for OMAP24xx UART2
  omap: Fix spotty MMC voltages
  ASoC: OMAP4: MCPDM: Remove unnecessary include of plat/control.h
  serial: omap-serial: fix signess error
  OMAP3: DMA: Errata i541: sDMA FIFO draining does not finish
  omap: dma: Fix buffering disable bit setting for omap24xx
  omap: serial: Fix the boot-up crash/reboot without CONFIG_PM
  OMAP3: PM: fix scratchpad memory accesses for off-mode
  omap4: pandaboard: enable the ehci port on pandaboard
  omap4: pandaboard: Fix the init if CONFIG_MMC_OMAP_HS is not set
  omap4: pandaboard: remove unused hsmmc definition
  OMAP: McBSP: Remove null omap44xx ops comment
  OMAP: McBSP: Swap CLKS source definition
  OMAP: McBSP: Fix CLKR and FSR signal muxing
  OMAP2+: clock: reduce the amount of standard debugging while disabling unused clocks
  OMAP: control: move plat-omap/control.h to mach-omap2/control.h
  OMAP: split plat-omap/common.c
  OMAP: McBSP: implement functional clock switching via clock framework
  OMAP: McBSP: implement McBSP CLKR and FSR signal muxing via mach-omap2/mcbsp.c
  ...

Fixed up trivial conflicts in arch/arm/mach-omap2/
{board-zoom-peripherals.c,devices.c} as per Tony