linux.git/mm/page_alloc.c, branch v4.13-rc2

mm, tree wide: replace __GFP_REPEAT by __GFP_RETRY_MAYFAIL with more useful semantic

2017-07-12T23:26:03+00:00

__GFP_REPEAT was designed to allow retry-but-eventually-fail semantic to
the page allocator.  This has been true but only for allocations
requests larger than PAGE_ALLOC_COSTLY_ORDER.  It has been always
ignored for smaller sizes.  This is a bit unfortunate because there is
no way to express the same semantic for those requests and they are
considered too important to fail so they might end up looping in the
page allocator for ever, similarly to GFP_NOFAIL requests.

Now that the whole tree has been cleaned up and accidental or misled
usage of __GFP_REPEAT flag has been removed for !costly requests we can
give the original flag a better name and more importantly a more useful
semantic.  Let's rename it to __GFP_RETRY_MAYFAIL which tells the user
that the allocator would try really hard but there is no promise of a
success.  This will work independent of the order and overrides the
default allocator behavior.  Page allocator users have several levels of
guarantee vs.  cost options (take GFP_KERNEL as an example)

 - GFP_KERNEL & ~__GFP_RECLAIM - optimistic allocation without _any_
   attempt to free memory at all. The most light weight mode which even
   doesn't kick the background reclaim. Should be used carefully because
   it might deplete the memory and the next user might hit the more
   aggressive reclaim

 - GFP_KERNEL & ~__GFP_DIRECT_RECLAIM (or GFP_NOWAIT)- optimistic
   allocation without any attempt to free memory from the current
   context but can wake kswapd to reclaim memory if the zone is below
   the low watermark. Can be used from either atomic contexts or when
   the request is a performance optimization and there is another
   fallback for a slow path.

 - (GFP_KERNEL|__GFP_HIGH) & ~__GFP_DIRECT_RECLAIM (aka GFP_ATOMIC) -
   non sleeping allocation with an expensive fallback so it can access
   some portion of memory reserves. Usually used from interrupt/bh
   context with an expensive slow path fallback.

 - GFP_KERNEL - both background and direct reclaim are allowed and the
   _default_ page allocator behavior is used. That means that !costly
   allocation requests are basically nofail but there is no guarantee of
   that behavior so failures have to be checked properly by callers
   (e.g. OOM killer victim is allowed to fail currently).

 - GFP_KERNEL | __GFP_NORETRY - overrides the default allocator behavior
   and all allocation requests fail early rather than cause disruptive
   reclaim (one round of reclaim in this implementation). The OOM killer
   is not invoked.

 - GFP_KERNEL | __GFP_RETRY_MAYFAIL - overrides the default allocator
   behavior and all allocation requests try really hard. The request
   will fail if the reclaim cannot make any progress. The OOM killer
   won't be triggered.

 - GFP_KERNEL | __GFP_NOFAIL - overrides the default allocator behavior
   and all allocation requests will loop endlessly until they succeed.
   This might be really dangerous especially for larger orders.

Existing users of __GFP_REPEAT are changed to __GFP_RETRY_MAYFAIL
because they already had their semantic.  No new users are added.
__alloc_pages_slowpath is changed to bail out for __GFP_RETRY_MAYFAIL if
there is no progress and we have already passed the OOM point.

This means that all the reclaim opportunities have been exhausted except
the most disruptive one (the OOM killer) and a user defined fallback
behavior is more sensible than keep retrying in the page allocator.

[akpm@linux-foundation.org: fix arch/sparc/kernel/mdesc.c]
[mhocko@suse.com: semantic fix]
  Link: http://lkml.kernel.org/r/20170626123847.GM11534@dhcp22.suse.cz
[mhocko@kernel.org: address other thing spotted by Vlastimil]
  Link: http://lkml.kernel.org/r/20170626124233.GN11534@dhcp22.suse.cz
Link: http://lkml.kernel.org/r/20170623085345.11304-3-mhocko@kernel.org
Signed-off-by: Michal Hocko 
Acked-by: Vlastimil Babka 
Cc: Alex Belits 
Cc: Chris Wilson 
Cc: Christoph Hellwig 
Cc: Darrick J. Wong 
Cc: David Daney 
Cc: Johannes Weiner 
Cc: Mel Gorman 
Cc: NeilBrown 
Cc: Ralf Baechle 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/memory-hotplug: switch locking to a percpu rwsem

2017-07-10T23:32:33+00:00

Andrey reported a potential deadlock with the memory hotplug lock and
the cpu hotplug lock.

The reason is that memory hotplug takes the memory hotplug lock and then
calls stop_machine() which calls get_online_cpus().  That's the reverse
lock order to get_online_cpus(); get_online_mems(); in mm/slub_common.c

The problem has been there forever.  The reason why this was never
reported is that the cpu hotplug locking had this homebrewn recursive
reader writer semaphore construct which due to the recursion evaded the
full lock dep coverage.  The memory hotplug code copied that construct
verbatim and therefor has similar issues.

Three steps to fix this:

1) Convert the memory hotplug locking to a per cpu rwsem so the
   potential issues get reported proper by lockdep.

2) Lock the online cpus in mem_hotplug_begin() before taking the memory
   hotplug rwsem and use stop_machine_cpuslocked() in the page_alloc
   code to avoid recursive locking.

3) The cpu hotpluck locking in #2 causes a recursive locking of the cpu
   hotplug lock via __offline_pages() -> lru_add_drain_all(). Solve this
   by invoking lru_add_drain_all_cpuslocked() instead.

Link: http://lkml.kernel.org/r/20170704093421.506836322@linutronix.de
Reported-by: Andrey Ryabinin 
Signed-off-by: Thomas Gleixner 
Acked-by: Michal Hocko 
Acked-by: Vlastimil Babka 
Cc: Vladimir Davydov 
Cc: Peter Zijlstra 
Cc: Davidlohr Bueso 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/page_alloc.c: eliminate unsigned confusion in __rmqueue_fallback

2017-07-10T23:32:32+00:00

Since current_order starts as MAX_ORDER-1 and is then only decremented,
the second half of the loop condition seems superfluous.  However, if
order is 0, we may decrement current_order past 0, making it UINT_MAX.
This is obviously too subtle ([1], [2]).

Since we need to add some comment anyway, change the two variables to
signed, making the counting-down for loop look more familiar, and
apparently also making gcc generate slightly smaller code.

[1] https://lkml.org/lkml/2016/6/20/493
[2] https://lkml.org/lkml/2017/6/19/345

[akpm@linux-foundation.org: fix up reject fixupping]
Link: http://lkml.kernel.org/r/20170621185529.2265-1-linux@rasmusvillemoes.dk
Signed-off-by: Rasmus Villemoes 
Reported-by: Hao Lee 
Acked-by: Wei Yang 
Acked-by: Michal Hocko 
Acked-by: Vlastimil Babka 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, page_alloc: fallback to smallest page when not stealing whole pageblock

2017-07-10T23:32:30+00:00

Since commit 3bc48f96cf11 ("mm, page_alloc: split smallest stolen page
in fallback") we pick the smallest (but sufficient) page of all that
have been stolen from a pageblock of different migratetype.  However,
there are cases when we decide not to steal the whole pageblock.

Practically in the current implementation it means that we are trying to
fallback for a MIGRATE_MOVABLE allocation of order X, go through the
freelists from MAX_ORDER-1 down to X, and find free page of order Y.  If
Y is less than pageblock_order / 2, we decide not to steal all pages
from the pageblock.  When Y > X, it means we are potentially splitting a
larger page than we need, as there might be other pages of order Z,
where X <= Z < Y.  Since Y is already too small to steal whole
pageblock, picking smallest available Z will result in the same decision
and we avoid splitting a higher-order page in a MIGRATE_UNMOVABLE or
MIGRATE_RECLAIMABLE pageblock.

This patch therefore changes the fallback algorithm so that in the
situation described above, we switch the fallback search strategy to go
from order X upwards to find the smallest suitable fallback.  In theory
there shouldn't be a downside of this change wrt fragmentation.

This has been tested with mmtests' stress-highalloc performing
GFP_KERNEL order-4 allocations, here is the relevant extfrag tracepoint
statistics:

                                                        4.12.0-rc2      4.12.0-rc2
                                                         1-kernel4       2-kernel4
  Page alloc extfrag event                                  25640976    69680977
  Extfrag fragmenting                                       25621086    69661364
  Extfrag fragmenting for unmovable                            74409       73204
  Extfrag fragmenting unmovable placed with movable            69003       67684
  Extfrag fragmenting unmovable placed with reclaim.            5406        5520
  Extfrag fragmenting for reclaimable                           6398        8467
  Extfrag fragmenting reclaimable placed with movable            869         884
  Extfrag fragmenting reclaimable placed with unmov.            5529        7583
  Extfrag fragmenting for movable                           25540279    69579693

Since we force movable allocations to steal the smallest available page
(which we then practially always split), we steal less per fallback, so
the number of fallbacks increases and steals potentially happen from
different pageblocks.  This is however not an issue for movable pages
that can be compacted.

Importantly, the "unmovable placed with movable" statistics is lower,
which is the result of less fragmentation in the unmovable pageblocks.
The effect on reclaimable allocation is a bit unclear.

Link: http://lkml.kernel.org/r/20170529093947.22618-1-vbabka@suse.cz
Signed-off-by: Vlastimil Babka 
Acked-by: Mel Gorman 
Cc: Johannes Weiner 
Cc: Joonsoo Kim 
Cc: David Rientjes 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, memory_hotplug: drop CONFIG_MOVABLE_NODE

2017-07-06T23:24:35+00:00

Commit 20b2f52b73fe ("numa: add CONFIG_MOVABLE_NODE for
movable-dedicated node") has introduced CONFIG_MOVABLE_NODE without a
good explanation on why it is actually useful.

It makes a lot of sense to make movable node semantic opt in but we
already have that because the feature has to be explicitly enabled on
the kernel command line.  A config option on top only makes the
configuration space larger without a good reason.  It also adds an
additional ifdefery that pollutes the code.

Just drop the config option and make it de-facto always enabled.  This
shouldn't introduce any change to the semantic.

Link: http://lkml.kernel.org/r/20170529114141.536-3-mhocko@kernel.org
Signed-off-by: Michal Hocko 
Acked-by: Reza Arbab 
Acked-by: Vlastimil Babka 
Cc: Mel Gorman 
Cc: Andrea Arcangeli 
Cc: Jerome Glisse 
Cc: Yasuaki Ishimatsu 
Cc: Xishi Qiu 
Cc: Kani Toshimitsu 
Cc: Chen Yucong 
Cc: Joonsoo Kim 
Cc: Andi Kleen 
Cc: David Rientjes 
Cc: Daniel Kiper 
Cc: Igor Mammedov 
Cc: Vitaly Kuznetsov 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm: vmstat: move slab statistics from zone to node counters

2017-07-06T23:24:35+00:00

Patch series "mm: per-lruvec slab stats"

Josef is working on a new approach to balancing slab caches and the page
cache.  For this to work, he needs slab cache statistics on the lruvec
level.  These patches implement that by adding infrastructure that
allows updating and reading generic VM stat items per lruvec, then
switches some existing VM accounting sites, including the slab
accounting ones, to this new cgroup-aware API.

I'll follow up with more patches on this, because there is actually
substantial simplification that can be done to the memory controller
when we replace private memcg accounting with making the existing VM
accounting sites cgroup-aware.  But this is enough for Josef to base his
slab reclaim work on, so here goes.

This patch (of 5):

To re-implement slab cache vs.  page cache balancing, we'll need the
slab counters at the lruvec level, which, ever since lru reclaim was
moved from the zone to the node, is the intersection of the node, not
the zone, and the memcg.

We could retain the per-zone counters for when the page allocator dumps
its memory information on failures, and have counters on both levels -
which on all but NUMA node 0 is usually redundant.  But let's keep it
simple for now and just move them.  If anybody complains we can restore
the per-zone counters.

[hannes@cmpxchg.org: fix oops]
  Link: http://lkml.kernel.org/r/20170605183511.GA8915@cmpxchg.org
Link: http://lkml.kernel.org/r/20170530181724.27197-3-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner 
Cc: Josef Bacik 
Cc: Michal Hocko 
Cc: Vladimir Davydov 
Cc: Rik van Riel 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, page_alloc: pass preferred nid instead of zonelist to allocator

2017-07-06T23:24:34+00:00

The main allocator function __alloc_pages_nodemask() takes a zonelist
pointer as one of its parameters.  All of its callers directly or
indirectly obtain the zonelist via node_zonelist() using a preferred
node id and gfp_mask.  We can make the code a bit simpler by doing the
zonelist lookup in __alloc_pages_nodemask(), passing it a preferred node
id instead (gfp_mask is already another parameter).

There are some code size benefits thanks to removal of inlined
node_zonelist():

  bloat-o-meter add/remove: 2/2 grow/shrink: 4/36 up/down: 399/-1351 (-952)

This will also make things simpler if we proceed with converting cpusets
to zonelists.

Link: http://lkml.kernel.org/r/20170517081140.30654-4-vbabka@suse.cz
Signed-off-by: Vlastimil Babka 
Reviewed-by: Christoph Lameter 
Acked-by: Michal Hocko 
Cc: Dimitri Sivanich 
Cc: "Kirill A. Shutemov" 
Cc: Andrea Arcangeli 
Cc: Anshuman Khandual 
Cc: David Rientjes 
Cc: Hugh Dickins 
Cc: Li Zefan 
Cc: Mel Gorman 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm, page_alloc: fix more premature OOM due to race with cpuset update

2017-07-06T23:24:34+00:00

I would like to stress that this patchset aims to fix issues and cleanup
the code *within the existing documented semantics*, i.e.  patch 1
ignores mempolicy restrictions if the set of allowed nodes has no
intersection with set of nodes allowed by cpuset.  I believe discussing
potential changes of the semantics can be better done once we have a
baseline with no known bugs of the current semantics.

I've recently summarized the cpuset/mempolicy issues in a LSF/MM
proposal [1] and the discussion itself [2].  I've been trying to rewrite
the handling as proposed, with the idea that changing semantics to make
all mempolicies static wrt cpuset updates (and discarding the relative
and default modes) can be tried on top, as there's a high risk of being
rejected/reverted because somebody might still care about the removed
modes.

However I haven't yet figured out how to properly:

1) make mempolicies swappable instead of rebinding in place. I thought
   mbind() already works that way and uses refcounting to avoid
   use-after-free of the old policy by a parallel allocation, but turns
   out true refcounting is only done for shared (shmem) mempolicies, and
   the actual protection for mbind() comes from mmap_sem. Extending the
   refcounting means more overhead in allocator hot path. Also swapping
   whole mempolicies means that we have to allocate the new ones, which
   can fail, and reverting of the partially done work also means
   allocating (note that mbind() doesn't care and will just leave part
   of the range updated and part not updated when returning -ENOMEM...).

2) make cpuset's task->mems_allowed also swappable (after converting it
   from nodemask to zonelist, which is the easy part) for mostly the
   same reasons.

The good news is that while trying to do the above, I've at least
figured out how to hopefully close the remaining premature OOM's, and do
a buch of cleanups on top, removing quite some of the code that was also
supposed to prevent the cpuset update races, but doesn't work anymore
nowadays.  This should fix the most pressing concerns with this topic
and give us a better baseline before either proceeding with the original
proposal, or pushing a change of semantics that removes the problem 1)
above.  I'd be then fine with trying to change the semantic first and
rewrite later.

Patchset has been tested with the LTP cpuset01 stress test.

[1] https://lkml.kernel.org/r/4c44a589-5fd8-08d0-892c-e893bb525b71@suse.cz
[2] https://lwn.net/Articles/717797/
[3] https://marc.info/?l=linux-mm&m=149191957922828&w=2

This patch (of 6):

Commit e47483bca2cc ("mm, page_alloc: fix premature OOM when racing with
cpuset mems update") has fixed known recent regressions found by LTP's
cpuset01 testcase.  I have however found that by modifying the testcase
to use per-vma mempolicies via bind(2) instead of per-task mempolicies
via set_mempolicy(2), the premature OOM still happens and the issue is
much older.

The root of the problem is that the cpuset's mems_allowed and
mempolicy's nodemask can temporarily have no intersection, thus
get_page_from_freelist() cannot find any usable zone.  The current
semantic for empty intersection is to ignore mempolicy's nodemask and
honour cpuset restrictions.  This is checked in node_zonelist(), but the
racy update can happen after we already passed the check.  Such races
should be protected by the seqlock task->mems_allowed_seq, but it
doesn't work here, because 1) mpol_rebind_mm() does not happen under
seqlock for write, and doing so would lead to deadlock, as it takes
mmap_sem for write, while the allocation can have mmap_sem for read when
it's taking the seqlock for read.  And 2) the seqlock cookie of callers
of node_zonelist() (alloc_pages_vma() and alloc_pages_current()) is
different than the one of __alloc_pages_slowpath(), so there's still a
potential race window.

This patch fixes the issue by having __alloc_pages_slowpath() check for
empty intersection of cpuset and ac->nodemask before OOM or allocation
failure.  If it's indeed empty, the nodemask is ignored and allocation
retried, which mimics node_zonelist().  This works fine, because almost
all callers of __alloc_pages_nodemask are obtaining the nodemask via
node_zonelist().  The only exception is new_node_page() from hotplug,
where the potential violation of nodemask isn't an issue, as there's
already a fallback allocation attempt without any nodemask.  If there's
a future caller that needs to have its specific nodemask honoured over
task's cpuset restrictions, we'll have to e.g.  add a gfp flag for that.

Link: http://lkml.kernel.org/r/20170517081140.30654-2-vbabka@suse.cz
Signed-off-by: Vlastimil Babka 
Acked-by: Michal Hocko 
Cc: Li Zefan 
Cc: Mel Gorman 
Cc: David Rientjes 
Cc: Christoph Lameter 
Cc: Hugh Dickins 
Cc: Andrea Arcangeli 
Cc: Anshuman Khandual 
Cc: "Kirill A. Shutemov" 
Cc: Dimitri Sivanich 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm/page_alloc.c: mark bad_range() and meminit_pfn_in_nid() as __maybe_unused

2017-07-06T23:24:33+00:00

The functions are not used in some configurations.  Adding the attribute
fixes the following warnings when building with clang:

  mm/page_alloc.c:409:19: error: function 'bad_range' is not needed and
      will not be emitted [-Werror,-Wunneeded-internal-declaration]

  mm/page_alloc.c:1106:30: error: unused function 'meminit_pfn_in_nid'
      [-Werror,-Wunused-function]

Link: http://lkml.kernel.org/r/20170518182030.165633-1-mka@chromium.org
Signed-off-by: Matthias Kaehlcke 
Cc: Vlastimil Babka 
Cc: Mel Gorman 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

mm: adaptive hash table scaling

2017-07-06T23:24:33+00:00

Allow hash tables to scale with memory but at slower pace, when
HASH_ADAPT is provided every time memory quadruples the sizes of hash
tables will only double instead of quadrupling as well.  This algorithm
starts working only when memory size reaches a certain point, currently
set to 64G.

This is example of dentry hash table size, before and after four various
memory configurations:

MEMORY	   SCALE	 HASH_SIZE
	old	new	old	new
    8G	 13	 13      8M      8M
   16G	 13	 13     16M     16M
   32G	 13	 13     32M     32M
   64G	 13	 13     64M     64M
  128G	 13	 14    128M     64M
  256G	 13	 14    256M    128M
  512G	 13	 15    512M    128M
 1024G	 13	 15   1024M    256M
 2048G	 13	 16   2048M    256M
 4096G	 13	 16   4096M    512M
 8192G	 13	 17   8192M    512M
16384G	 13	 17  16384M   1024M
32768G	 13	 18  32768M   1024M
65536G	 13	 18  65536M   2048M

The effect of this change on runtime is undetectable as filesystem
growth is not proportional to machine memory size as is currently
assumed.  The change effects only large memory machine.  Additional
tuning might be needed, but that can be done by the clients of the
kmem_cache_create interface, not the generic cache allocator itself.

The adaptive hashing is disabled on 32 bit systems to avoid confusion of
whether base should be different for smaller systems, and to avoid
overflows.

[mhocko@suse.com: drop HASH_ADAPT]
  Link: http://lkml.kernel.org/r/20170509094607.GG6481@dhcp22.suse.cz
[pasha.tatashin@oracle.com: UL -> ULL fix]
  Link: http://lkml.kernel.org/r/1495300013-653283-2-git-send-email-pasha.tatashin@oracle.com
[pasha.tatashin@oracle.com: disable adaptive hash on 32 bit systems]
  Link: http://lkml.kernel.org/r/1495469329-755807-2-git-send-email-pasha.tatashin@oracle.com
Link: http://lkml.kernel.org/r/1488432825-92126-5-git-send-email-pasha.tatashin@oracle.com
Signed-off-by: Pavel Tatashin 
Signed-off-by: Michal Hocko 
Acked-by: Michal Hocko 
Cc: David Miller 
Cc: Al Viro 
Cc: Babu Moger 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds