linux.git/mm/compaction.c, branch v4.6-rc2

mm, kswapd: replace kswapd compaction with waking up kcompactd

2016-03-17T22:09:34+00:00

Similarly to direct reclaim/compaction, kswapd attempts to combine
reclaim and compaction to attempt making memory allocation of given
order available.

The details differ from direct reclaim e.g. in having high watermark as
a goal.  The code involved in kswapd's reclaim/compaction decisions has
evolved to be quite complex.

Testing reveals that it doesn't actually work in at least one scenario,
and closer inspection suggests that it could be greatly simplified
without compromising on the goal (make high-order page available) or
efficiency (don't reclaim too much).  The simplification relieas of
doing all compaction in kcompactd, which is simply woken up when high
watermarks are reached by kswapd's reclaim.

The scenario where kswapd compaction doesn't work was found with mmtests
test stress-highalloc configured to attempt order-9 allocations without
direct reclaim, just waking up kswapd.  There was no compaction attempt
from kswapd during the whole test.  Some added instrumentation shows
what happens:

 - balance_pgdat() sets end_zone to Normal, as it's not balanced
 - reclaim is attempted on DMA zone, which sets nr_attempted to 99, but
   it cannot reclaim anything, so sc.nr_reclaimed is 0
 - for zones DMA32 and Normal, kswapd_shrink_zone uses testorder=0, so
   it merely checks if high watermarks were reached for base pages.
   This is true, so no reclaim is attempted.  For DMA, testorder=0
   wasn't used, as compaction_suitable() returned COMPACT_SKIPPED
 - even though the pgdat_needs_compaction flag wasn't set to false, no
   compaction happens due to the condition sc.nr_reclaimed >
   nr_attempted being false (as 0 < 99)
 - priority-- due to nr_reclaimed being 0, repeat until priority reaches
   0 pgdat_balanced() is false as only the small zone DMA appears
   balanced (curiously in that check, watermark appears OK and
   compaction_suitable() returns COMPACT_PARTIAL, because a lower
   classzone_idx is used there)

Now, even if it was decided that reclaim shouldn't be attempted on the
DMA zone, the scenario would be the same, as (sc.nr_reclaimed=0 >
nr_attempted=0) is also false.  The condition really should use >= as
the comment suggests.  Then there is a mismatch in the check for setting
pgdat_needs_compaction to false using low watermark, while the rest uses
high watermark, and who knows what other subtlety.  Hopefully this
demonstrates that this is unsustainable.

Luckily we can simplify this a lot.  The reclaim/compaction decisions
make sense for direct reclaim scenario, but in kswapd, our primary goal
is to reach high watermark in order-0 pages.  Afterwards we can attempt
compaction just once.  Unlike direct reclaim, we don't reclaim extra
pages (over the high watermark), the current code already disallows it
for good reasons.

After this patch, we simply wake up kcompactd to process the pgdat,
after we have either succeeded or failed to reach the high watermarks in
kswapd, which goes to sleep.  We pass kswapd's order and classzone_idx,
so kcompactd can apply the same criteria to determine which zones are
worth compacting.  Note that we use the classzone_idx from
wakeup_kswapd(), not balanced_classzone_idx which can include higher
zones that kswapd tried to balance too, but didn't consider them in
pgdat_balanced().

Since kswapd now cannot create high-order pages itself, we need to
adjust how it determines the zones to be balanced.  The key element here
is adding a "highorder" parameter to zone_balanced, which, when set to
false, makes it consider only order-0 watermark instead of the desired
higher order (this was done previously by kswapd_shrink_zone(), but not
elsewhere).  This false is passed for example in pgdat_balanced().
Importantly, wakeup_kswapd() uses true to make sure kswapd and thus
kcompactd are woken up for a high-order allocation failure.

The last thing is to decide what to do with pageblock_skip bitmap
handling.  Compaction maintains a pageblock_skip bitmap to record
pageblocks where isolation recently failed.  This bitmap can be reset by
three ways:

1) direct compaction is restarting after going through the full deferred cycle

2) kswapd goes to sleep, and some other direct compaction has previously
   finished scanning the whole zone and set zone->compact_blockskip_flush.
   Note that a successful direct compaction clears this flag.

3) compaction was invoked manually via trigger in /proc

The case 2) is somewhat fuzzy to begin with, but after introducing
kcompactd we should update it.  The check for direct compaction in 1),
and to set the flush flag in 2) use current_is_kswapd(), which doesn't
work for kcompactd.  Thus, this patch adds bool direct_compaction to
compact_control to use in 2).  For the case 1) we remove the check
completely - unlike the former kswapd compaction, kcompactd does use the
deferred compaction functionality, so flushing tied to restarting from
deferred compaction makes sense here.

Note that when kswapd goes to sleep, kcompactd is woken up, so it will
see the flushed pageblock_skip bits.  This is different from when the
former kswapd compaction observed the bits and I believe it makes more
sense.  Kcompactd can afford to be more thorough than a direct
compaction trying to limit allocation latency, or kswapd whose primary
goal is to reclaim.

For testing, I used stress-highalloc configured to do order-9
allocations with GFP_NOWAIT|__GFP_HIGH|__GFP_COMP, so they relied just
on kswapd/kcompactd reclaim/compaction (the interfering kernel builds in
phases 1 and 2 work as usual):

stress-highalloc
                        4.5-rc1+before          4.5-rc1+after
                             -nodirect              -nodirect
Success 1 Min          1.00 (  0.00%)         5.00 (-66.67%)
Success 1 Mean         1.40 (  0.00%)         6.20 (-55.00%)
Success 1 Max          2.00 (  0.00%)         7.00 (-16.67%)
Success 2 Min          1.00 (  0.00%)         5.00 (-66.67%)
Success 2 Mean         1.80 (  0.00%)         6.40 (-52.38%)
Success 2 Max          3.00 (  0.00%)         7.00 (-16.67%)
Success 3 Min         34.00 (  0.00%)        62.00 (  1.59%)
Success 3 Mean        41.80 (  0.00%)        63.80 (  1.24%)
Success 3 Max         53.00 (  0.00%)        65.00 (  2.99%)

User                          3166.67        3181.09
System                        1153.37        1158.25
Elapsed                       1768.53        1799.37

                            4.5-rc1+before   4.5-rc1+after
                                 -nodirect    -nodirect
Direct pages scanned                32938        32797
Kswapd pages scanned              2183166      2202613
Kswapd pages reclaimed            2152359      2143524
Direct pages reclaimed              32735        32545
Percentage direct scans                1%           1%
THP fault alloc                       579          612
THP collapse alloc                    304          316
THP splits                              0            0
THP fault fallback                    793          778
THP collapse fail                      11           16
Compaction stalls                    1013         1007
Compaction success                     92           67
Compaction failures                   920          939
Page migrate success               238457       721374
Page migrate failure                23021        23469
Compaction pages isolated          504695      1479924
Compaction migrate scanned         661390      8812554
Compaction free scanned          13476658     84327916
Compaction cost                       262          838

After this patch we see improvements in allocation success rate
(especially for phase 3) along with increased compaction activity.  The
compaction stalls (direct compaction) in the interfering kernel builds
(probably THP's) also decreased somewhat thanks to kcompactd activity,
yet THP alloc successes improved a bit.

Note that elapsed and user time isn't so useful for this benchmark,
because of the background interference being unpredictable.  It's just
to quickly spot some major unexpected differences.  System time is
somewhat more useful and that didn't increase.

Also (after adjusting mmtests' ftrace monitor):

Time kswapd awake               2547781     2269241
Time kcompactd awake                  0      119253
Time direct compacting           939937      557649
Time kswapd compacting                0           0
Time kcompactd compacting             0      119099

The decrease of overal time spent compacting appears to not match the
increased compaction stats.  I suspect the tasks get rescheduled and
since the ftrace monitor doesn't see that, the reported time is wall
time, not CPU time.  But arguably direct compactors care about overall
latency anyway, whether busy compacting or waiting for CPU doesn't
matter.  And that latency seems to almost halved.

It's also interesting how much time kswapd spent awake just going
through all the priorities and failing to even try compacting, over and
over.

We can also configure stress-highalloc to perform both direct
reclaim/compaction and wakeup kswapd/kcompactd, by using
GFP_KERNEL|__GFP_HIGH|__GFP_COMP:

stress-highalloc
                        4.5-rc1+before         4.5-rc1+after
                               -direct               -direct
Success 1 Min          4.00 (  0.00%)        9.00 (-50.00%)
Success 1 Mean         8.00 (  0.00%)       10.00 (-19.05%)
Success 1 Max         12.00 (  0.00%)       11.00 ( 15.38%)
Success 2 Min          4.00 (  0.00%)        9.00 (-50.00%)
Success 2 Mean         8.20 (  0.00%)       10.00 (-16.28%)
Success 2 Max         13.00 (  0.00%)       11.00 (  8.33%)
Success 3 Min         75.00 (  0.00%)       74.00 (  1.33%)
Success 3 Mean        75.60 (  0.00%)       75.20 (  0.53%)
Success 3 Max         77.00 (  0.00%)       76.00 (  0.00%)

User                          3344.73       3246.04
System                        1194.24       1172.29
Elapsed                       1838.04       1836.76

                            4.5-rc1+before  4.5-rc1+after
                                   -direct     -direct
Direct pages scanned               125146      120966
Kswapd pages scanned              2119757     2135012
Kswapd pages reclaimed            2073183     2108388
Direct pages reclaimed             124909      120577
Percentage direct scans                5%          5%
THP fault alloc                       599         652
THP collapse alloc                    323         354
THP splits                              0           0
THP fault fallback                    806         793
THP collapse fail                      17          16
Compaction stalls                    2457        2025
Compaction success                    906         518
Compaction failures                  1551        1507
Page migrate success              2031423     2360608
Page migrate failure                32845       40852
Compaction pages isolated         4129761     4802025
Compaction migrate scanned       11996712    21750613
Compaction free scanned         214970969   344372001
Compaction cost                      2271        2694

In this scenario, this patch doesn't change the overall success rate as
direct compaction already tries all it can.  There's however significant
reduction in direct compaction stalls (that is, the number of
allocations that went into direct compaction).  The number of successes
(i.e.  direct compaction stalls that ended up with successful
allocation) is reduced by the same number.  This means the offload to
kcompactd is working as expected, and direct compaction is reduced
either due to detecting contention, or compaction deferred by kcompactd.
In the previous version of this patchset there was some apparent
reduction of success rate, but the changes in this version (such as
using sync compaction only), new baseline kernel, and/or averaging
results from 5 executions (my bet), made this go away.

Ftrace-based stats seem to roughly agree:

Time kswapd awake               2532984     2326824
Time kcompactd awake                  0      257916
Time direct compacting           864839      735130
Time kswapd compacting                0           0
Time kcompactd compacting             0      257585

Signed-off-by: Vlastimil Babka 
Cc: Andrea Arcangeli 
Cc: "Kirill A. Shutemov" 
Cc: Rik van Riel 
Cc: Joonsoo Kim 
Cc: Mel Gorman 
Cc: David Rientjes 
Cc: Michal Hocko 
Cc: Johannes Weiner 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, compaction: introduce kcompactd

2016-03-17T22:09:34+00:00

Memory compaction can be currently performed in several contexts:

 - kswapd balancing a zone after a high-order allocation failure
 - direct compaction to satisfy a high-order allocation, including THP
   page fault attemps
 - khugepaged trying to collapse a hugepage
 - manually from /proc

The purpose of compaction is two-fold.  The obvious purpose is to
satisfy a (pending or future) high-order allocation, and is easy to
evaluate.  The other purpose is to keep overal memory fragmentation low
and help the anti-fragmentation mechanism.  The success wrt the latter
purpose is more

The current situation wrt the purposes has a few drawbacks:

 - compaction is invoked only when a high-order page or hugepage is not
   available (or manually).  This might be too late for the purposes of
   keeping memory fragmentation low.
 - direct compaction increases latency of allocations.  Again, it would
   be better if compaction was performed asynchronously to keep
   fragmentation low, before the allocation itself comes.
 - (a special case of the previous) the cost of compaction during THP
   page faults can easily offset the benefits of THP.
 - kswapd compaction appears to be complex, fragile and not working in
   some scenarios.  It could also end up compacting for a high-order
   allocation request when it should be reclaiming memory for a later
   order-0 request.

To improve the situation, we should be able to benefit from an
equivalent of kswapd, but for compaction - i.e. a background thread
which responds to fragmentation and the need for high-order allocations
(including hugepages) somewhat proactively.

One possibility is to extend the responsibilities of kswapd, which could
however complicate its design too much.  It should be better to let
kswapd handle reclaim, as order-0 allocations are often more critical
than high-order ones.

Another possibility is to extend khugepaged, but this kthread is a
single instance and tied to THP configs.

This patch goes with the option of a new set of per-node kthreads called
kcompactd, and lays the foundations, without introducing any new
tunables.  The lifecycle mimics kswapd kthreads, including the memory
hotplug hooks.

For compaction, kcompactd uses the standard compaction_suitable() and
ompact_finished() criteria and the deferred compaction functionality.
Unlike direct compaction, it uses only sync compaction, as there's no
allocation latency to minimize.

This patch doesn't yet add a call to wakeup_kcompactd.  The kswapd
compact/reclaim loop for high-order pages will be replaced by waking up
kcompactd in the next patch with the description of what's wrong with
the old approach.

Waking up of the kcompactd threads is also tied to kswapd activity and
follows these rules:
 - we don't want to affect any fastpaths, so wake up kcompactd only from
   the slowpath, as it's done for kswapd
 - if kswapd is doing reclaim, it's more important than compaction, so
   don't invoke kcompactd until kswapd goes to sleep
 - the target order used for kswapd is passed to kcompactd

Future possible future uses for kcompactd include the ability to wake up
kcompactd on demand in special situations, such as when hugepages are
not available (currently not done due to __GFP_NO_KSWAPD) or when a
fragmentation event (i.e.  __rmqueue_fallback()) occurs.  It's also
possible to perform periodic compaction with kcompactd.

[arnd@arndb.de: fix build errors with kcompactd]
[paul.gortmaker@windriver.com: don't use modular references for non modular code]
Signed-off-by: Vlastimil Babka 
Cc: Andrea Arcangeli 
Cc: "Kirill A. Shutemov" 
Cc: Rik van Riel 
Cc: Joonsoo Kim 
Cc: Mel Gorman 
Cc: David Rientjes 
Cc: Michal Hocko 
Cc: Johannes Weiner 
Signed-off-by: Arnd Bergmann 
Signed-off-by: Paul Gortmaker 
Cc: Hugh Dickins 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

2016-03-15T23:55:16+00:00

There is a performance drop report due to hugepage allocation and in
there half of cpu time are spent on pageblock_pfn_to_page() in
compaction [1].

In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails.  Most costly operations in this
case is to find valid pageblock while scanning whole zone range.  To
check if pageblock is valid to compact, valid pfn within pageblock is
required and we can obtain it by calling pageblock_pfn_to_page().  This
function checks whether pageblock is in a single zone and return valid
pfn if possible.  Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to be very
expensive in this workload.

Although we have no way to skip this pageblock check in the system where
hole exists at arbitrary position, we can use cached value for zone
continuity and just do pfn_to_page() in the system where hole doesn't
exist.  This optimization considerably speeds up in above workload.

Before vs After
  Max: 1096 MB/s vs 1325 MB/s
  Min: 635 MB/s 1015 MB/s
  Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

[akpm@linux-foundation.org: don't forget to restore zone->contiguous on error path, per Vlastimil]
Signed-off-by: Joonsoo Kim 
Reported-by: Aaron Lu 
Acked-by: Vlastimil Babka 
Tested-by: Aaron Lu 
Cc: Mel Gorman 
Cc: Rik van Riel 
Cc: David Rientjes 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/compaction: pass only pageblock aligned range to pageblock_pfn_to_page

2016-03-15T23:55:16+00:00

pageblock_pfn_to_page() is used to check there is valid pfn and all
pages in the pageblock is in a single zone.  If there is a hole in the
pageblock, passing arbitrary position to pageblock_pfn_to_page() could
cause to skip whole pageblock scanning, instead of just skipping the
hole page.  For deterministic behaviour, it's better to always pass
pageblock aligned range to pageblock_pfn_to_page().  It will also help
further optimization on pageblock_pfn_to_page() in the following patch.

Signed-off-by: Joonsoo Kim 
Cc: Aaron Lu 
Cc: David Rientjes 
Cc: Mel Gorman 
Cc: Rik van Riel 
Acked-by: Vlastimil Babka 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/compaction: fix invalid free_pfn and compact_cached_free_pfn

2016-03-15T23:55:16+00:00

free_pfn and compact_cached_free_pfn are the pointer that remember
restart position of freepage scanner.  When they are reset or invalid,
we set them to zone_end_pfn because freepage scanner works in reverse
direction.  But, because zone range is defined as [zone_start_pfn,
zone_end_pfn), zone_end_pfn is invalid to access.  Therefore, we should
not store it to free_pfn and compact_cached_free_pfn.  Instead, we need
to store zone_end_pfn - 1 to them.  There is one more thing we should
consider.  Freepage scanner scan reversely by pageblock unit.  If
free_pfn and compact_cached_free_pfn are set to middle of pageblock, it
regards that sitiation as that it already scans front part of pageblock
so we lose opportunity to scan there.  To fix-up, this patch do
round_down() to guarantee that reset position will be pageblock aligned.

Note that thanks to the current pageblock_pfn_to_page() implementation,
actual access to zone_end_pfn doesn't happen until now.  But, following
patch will change pageblock_pfn_to_page() so this patch is needed from
now on.

Signed-off-by: Joonsoo Kim 
Acked-by: David Rientjes 
Acked-by: Vlastimil Babka 
Cc: Aaron Lu 
Cc: Mel Gorman 
Cc: Rik van Riel 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/compaction.c: __compact_pgdat() code cleanuup

2016-01-15T00:00:49+00:00

This patch uses is_via_compact_memory() to distinguish compaction from
sysfs or sysctl.  And, this patch also reduces indentation on
compaction_defer_reset() by filtering these cases first before checking
watermark.

There is no functional change.

Signed-off-by: Joonsoo Kim 
Acked-by: Yaowei Bai 
Acked-by: David Rientjes 
Acked-by: Vlastimil Babka 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/compaction: improve comment for compact_memory tunable knob handler

2016-01-15T00:00:49+00:00

sysctl_compaction_handler() is the handler function for compact_memory
tunable knob under /proc/sys/vm, add the missing knob name to make this
more accurate in comment.

No functional change.

Signed-off-by: Yaowei Bai 
Acked-by: Vlastimil Babka 
Acked-by: Michal Nazarewicz 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, compaction: distinguish contended status in tracepoints

2015-11-06T03:34:48+00:00

Compaction returns prematurely with COMPACT_PARTIAL when contended or has
fatal signal pending.  This is ok for the callers, but might be misleading
in the traces, as the usual reason to return COMPACT_PARTIAL is that we
think the allocation should succeed.  After this patch we distinguish the
premature ending condition in the mm_compaction_finished and
mm_compaction_end tracepoints.

The contended status covers the following reasons:
- lock contention or need_resched() detected in async compaction
- fatal signal pending
- too many pages isolated in the zone (only for async compaction)
Further distinguishing the exact reason seems unnecessary for now.

Signed-off-by: Vlastimil Babka 
Cc: Joonsoo Kim 
Cc: Mel Gorman 
Cc: David Rientjes 
Cc: Steven Rostedt 
Cc: Ingo Molnar 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, compaction: export tracepoints status strings to userspace

2015-11-06T03:34:48+00:00

Some compaction tracepoints convert the integer return values to strings
using the compaction_status_string array.  This works for in-kernel
printing, but not userspace trace printing of raw captured trace such as
via trace-cmd report.

This patch converts the private array to appropriate tracepoint macros
that result in proper userspace support.

trace-cmd output before:
transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
  zone=ffffffff81815d7a order=9 ret=

after:
transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
  zone=ffffffff81815d7a order=9 ret=partial

Signed-off-by: Vlastimil Babka 
Reviewed-by: Steven Rostedt 
Cc: Joonsoo Kim 
Cc: Ingo Molnar 
Cc: Mel Gorman 
Cc: David Rientjes 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/compaction.c: add an is_via_compact_memory() helper

2015-11-06T03:34:48+00:00

Introduce is_via_compact_memory() helper indicating compacting via
/proc/sys/vm/compact_memory to improve readability.

To catch this situation in __compaction_suitable, use order as parameter
directly instead of using struct compact_control.

This patch has no functional changes.

Signed-off-by: Yaowei Bai 
Cc: Mel Gorman 
Acked-by: David Rientjes 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds