Add debug checks for invariants that if broken could lead to mapcount vs
page_mapcount debug checks to trigger later in split_huge_page.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add hugepage statistics to per-node sysfs meminfo
Reviewed-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NDavid Rientjes <rientjes@google.com>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Make sure we scale up nr_rotated when we encounter a referenced
transparent huge page.  This ensures pageout scanning balance is not
distorted when there are huge pages on the LRU.
Signed-off-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Count each transparent hugepage as HPAGE_PMD_NR pages in the LRU
statistics, so the Active(anon) and Inactive(anon) statistics in
/proc/meminfo are correct.
Signed-off-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

On small systems, the extra memory used by the anti-fragmentation memory
reserve and simply because huge pages are smaller than large pages can
easily outweigh the benefits of less TLB misses.

A less obvious concern is if run on a NUMA machine with asymmetric node
sizes and one of them is very small.  The reserve could make the node
unusable.

In case of the crashdump kernel, OOMs have been observed due to the
anti-fragmentation memory reserve taking up a large fraction of the
crashdump image.

This patch disables transparent hugepages on systems with less than 1GB of
RAM, but the hugepage subsystem is fully initialized so administrators can
enable THP through /sys if desired.
Signed-off-by: NRik van Riel <riel@redhat.com>
Acked-by: NAvi Kiviti <avi@redhat.com>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It makes no sense not to enable compaction for small order pages as we
don't want to end up with bad order 2 allocations and good and graceful
order 9 allocations.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This takes advantage of memory compaction to properly generate pages of
order > 0 if regular page reclaim fails and priority level becomes more
severe and we don't reach the proper watermarks.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It's unclear why schedule friendly kernel threads can't be taken away by
the CPU through the scheduler itself.  It's safer to stop them as they can
trigger memory allocation, if kswapd also freezes itself to avoid
generating I/O they have too.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

For GRU and EPT, we need gup-fast to set referenced bit too (this is why
it's correct to return 0 when shadow_access_mask is zero, it requires
gup-fast to set the referenced bit).  qemu-kvm access already sets the
young bit in the pte if it isn't zero-copy, if it's zero copy or a shadow
paging EPT minor fault we relay on gup-fast to signal the page is in
use...

We also need to check the young bits on the secondary pagetables for NPT
and not nested shadow mmu as the data may never get accessed again by the
primary pte.

Without this closer accuracy, we'd have to remove the heuristic that
avoids collapsing hugepages in hugepage virtual regions that have not even
a single subpage in use.

->test_young is full backwards compatible with GRU and other usages that
don't have young bits in pagetables set by the hardware and that should
nuke the secondary mmu mappings when ->clear_flush_young runs just like
EPT does.

Removing the heuristic that checks the young bit in
khugepaged/collapse_huge_page completely isn't so bad either probably but
I thought it was worth it and this makes it reliable.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Archs implementing Transparent Hugepage Support must implement a function
called has_transparent_hugepage to be sure the virtual or physical CPU
supports Transparent Hugepages.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

An huge pmd can only be mapped if the corresponding 2M virtual range is
fully contained in the vma.  At times the VM calls split_vma twice, if the
first split_vma succeeds and the second fail, the first split_vma remains
in effect and it's not rolled back.  For split_vma or vma_adjust to fail
an allocation failure is needed so it's a very unlikely event (the out of
memory killer would normally fire before any allocation failure is visible
to kernel and userland and if an out of memory condition happens it's
unlikely to happen exactly here).  Nevertheless it's safer to ensure that
no huge pmd can be left around if the vma is adjusted in a way that can't
fit hugepages anymore at the new vm_start/vm_end address.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It's not worth migrating transparent hugepages during compaction.  Those
hugepages don't create fragmentation.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

With transparent hugepage support we need compaction for the "defrag"
sysfs controls to be effective.

At the moment THP hangs the system if COMPACTION isn't selected, as
without COMPACTION lumpy reclaim wouldn't be entirely disabled.  So at the
moment it's not orthogonal.  When lumpy will be removed from the VM I can
remove the select COMPACTION in theory, but then 99% of THP users would be
still doing a mistake in disabling compaction, even if the mistake won't
return in fatal runtime but just slightly degraded performance.  So from a
theoretical standpoing forcing the below select is not needed (the
dependency isn't strict nor at compile time nor at runtime) but from a
practical standpoint it is safer.

If anybody really wants THP to run without compaction, it'd be such a
weird setup that editing the Kconfig file to allow it will be surely not a
problem.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Allow to choose between the always|madvise default for page faults and
khugepaged at config time.  madvise guarantees zero risk of higher memory
footprint for applications (applications using madvise(MADV_HUGEPAGE)
won't risk to use any more memory by backing their virtual regions with
hugepages).

Initially set the default to N and don't depend on EMBEDDED.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This tries to be more friendly to filesystem in userland, with userland
backends that allocate memory in the I/O paths and that could deadlock if
khugepaged holds the mmap_sem write mode of the userland backend while
allocating memory.  Memory allocation may wait for writeback I/O
completion from the daemon that may be blocked in the mmap_sem read mode
if a page fault happens and the daemon wasn't using mlock for the memory
required for the I/O submission and completion.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It's mostly a matter of replacing alloc_pages with alloc_pages_vma after
introducing alloc_pages_vma.  khugepaged needs special handling as the
allocation has to happen inside collapse_huge_page where the vma is known
and an error has to be returned to the outer loop to sleep
alloc_sleep_millisecs in case of failure.  But it retains the more
efficient logic of handling allocation failures in khugepaged in case of
CONFIG_NUMA=n.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

With memory compaction in, and lumpy-reclaim disabled, it seems safe
enough to defrag memory during the (synchronous) transparent hugepage page
faults (TRANSPARENT_HUGEPAGE_DEFRAG_FLAG) and not only during khugepaged
(async) hugepage allocations that was already enabled even before memory
compaction was in (TRANSPARENT_HUGEPAGE_DEFRAG_KHUGEPAGED_FLAG).
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

If transparent hugepage is enabled initialize min_free_kbytes to an
optimal value by default.  This moves the hugeadm algorithm in kernel.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Natively handle huge pmds when changing page tables on behalf of
mprotect().

I left out update_mmu_cache() because we do not need it on x86 anyway but
more importantly the interface works on ptes, not pmds.
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Flushing the tlb for huge pmds requires the vma's anon_vma, so pass along
the vma instead of the mm, we can always get the latter when we need it.
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add pmd_modify() for use with mprotect() on huge pmds.
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Handle transparent huge page pmd entries natively instead of splitting
them into subpages.
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add support for transparent hugepages to x86 32bit.

Share the same VM_ bitflag for VM_MAPPED_COPY.  mm/nommu.c will never
support transparent hugepages.
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

PG_buddy can be converted to _mapcount == -2.  So the PG_compound_lock can
be added to page->flags without overflowing (because of the sparse section
bits increasing) with CONFIG_X86_PAE=y and CONFIG_X86_PAT=y.  This also
has to move the memory hotplug code from _mapcount to lru.next to avoid
any risk of clashes.  We can't use lru.next for PG_buddy removal, but
memory hotplug can use lru.next even more easily than the mapcount
instead.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Skip transhuge pages in ksm for now.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

register in khugepaged if the vma grows.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add khugepaged to relocate fragmented pages into hugepages if new
hugepages become available.  (this is indipendent of the defrag logic that
will have to make new hugepages available)

The fundamental reason why khugepaged is unavoidable, is that some memory
can be fragmented and not everything can be relocated.  So when a virtual
machine quits and releases gigabytes of hugepages, we want to use those
freely available hugepages to create huge-pmd in the other virtual
machines that may be running on fragmented memory, to maximize the CPU
efficiency at all times.  The scan is slow, it takes nearly zero cpu time,
except when it copies data (in which case it means we definitely want to
pay for that cpu time) so it seems a good tradeoff.

In addition to the hugepages being released by other process releasing
memory, we have the strong suspicion that the performance impact of
potentially defragmenting hugepages during or before each page fault could
lead to more performance inconsistency than allocating small pages at
first and having them collapsed into large pages later...  if they prove
themselfs to be long lived mappings (khugepaged scan is slow so short
lived mappings have low probability to run into khugepaged if compared to
long lived mappings).
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add hugepage stat information to /proc/vmstat and /proc/meminfo.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add memcg charge/uncharge to hugepage faults in huge_memory.c.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

By this patch, when a transparent hugepage is charged, not only the head
page but also all the tail pages are committed, IOW pc->mem_cgroup and
pc->flags of tail pages are set.

Without this patch:

- Tail pages are not linked to any memcg's LRU at splitting. This causes many
  problems, for example, the charged memcg's directory can never be rmdir'ed
  because it doesn't have enough pages to scan to make the usage decrease to 0.
- "rss" field in memory.stat would be incorrect. Moreover, usage_in_bytes in
  root cgroup is calculated by the stat not by res_counter(since 2.6.32),
  it would be incorrect too.
Signed-off-by: NDaisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Teach memcg to charge/uncharge compound pages.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

No pmd_trans_huge should ever materialize in migration ptes areas, because
we split the hugepage before migration ptes are instantiated.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add madvise MADV_HUGEPAGE to mark regions that are important to be
hugepage backed.  Return -EINVAL if the vma is not of an anonymous type,
or the feature isn't built into the kernel.  Never silently return
success.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

pte_trans_huge must not leak in certain vmas like the mmio special pfn or
filebacked mappings.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This documents how split_huge_page is safe vs new vma inserctions into the
anon_vma that may have already released the anon_vma->lock but not
established pmds yet when split_huge_page starts.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

If you configure THP in addition to HUGETLB_PAGE on x86_32 without PAE,
the p?d-folding works out that munlock_vma_pages_range() can crash to
follow_page()'s pud_huge() BUG_ON(flags & FOLL_GET): it needs the same
VM_HUGETLB check already there on the pmd_huge() line.  Conveniently,
openSUSE provides a "blogd" which tests this out at startup!
Signed-off-by: NHugh Dickins <hughd@google.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Lately I've been working to make KVM use hugepages transparently without
the usual restrictions of hugetlbfs.  Some of the restrictions I'd like to
see removed:

1) hugepages have to be swappable or the guest physical memory remains
   locked in RAM and can't be paged out to swap

2) if a hugepage allocation fails, regular pages should be allocated
   instead and mixed in the same vma without any failure and without
   userland noticing

3) if some task quits and more hugepages become available in the
   buddy, guest physical memory backed by regular pages should be
   relocated on hugepages automatically in regions under
   madvise(MADV_HUGEPAGE) (ideally event driven by waking up the
   kernel deamon if the order=HPAGE_PMD_SHIFT-PAGE_SHIFT list becomes
   not null)

4) avoidance of reservation and maximization of use of hugepages whenever
   possible. Reservation (needed to avoid runtime fatal faliures) may be ok for
   1 machine with 1 database with 1 database cache with 1 database cache size
   known at boot time. It's definitely not feasible with a virtualization
   hypervisor usage like RHEV-H that runs an unknown number of virtual machines
   with an unknown size of each virtual machine with an unknown amount of
   pagecache that could be potentially useful in the host for guest not using
   O_DIRECT (aka cache=off).

hugepages in the virtualization hypervisor (and also in the guest!) are
much more important than in a regular host not using virtualization,
becasue with NPT/EPT they decrease the tlb-miss cacheline accesses from 24
to 19 in case only the hypervisor uses transparent hugepages, and they
decrease the tlb-miss cacheline accesses from 19 to 15 in case both the
linux hypervisor and the linux guest both uses this patch (though the
guest will limit the addition speedup to anonymous regions only for
now...).  Even more important is that the tlb miss handler is much slower
on a NPT/EPT guest than for a regular shadow paging or no-virtualization
scenario.  So maximizing the amount of virtual memory cached by the TLB
pays off significantly more with NPT/EPT than without (even if there would
be no significant speedup in the tlb-miss runtime).

The first (and more tedious) part of this work requires allowing the VM to
handle anonymous hugepages mixed with regular pages transparently on
regular anonymous vmas.  This is what this patch tries to achieve in the
least intrusive possible way.  We want hugepages and hugetlb to be used in
a way so that all applications can benefit without changes (as usual we
leverage the KVM virtualization design: by improving the Linux VM at
large, KVM gets the performance boost too).

The most important design choice is: always fallback to 4k allocation if
the hugepage allocation fails!  This is the _very_ opposite of some large
pagecache patches that failed with -EIO back then if a 64k (or similar)
allocation failed...

Second important decision (to reduce the impact of the feature on the
existing pagetable handling code) is that at any time we can split an
hugepage into 512 regular pages and it has to be done with an operation
that can't fail.  This way the reliability of the swapping isn't decreased
(no need to allocate memory when we are short on memory to swap) and it's
trivial to plug a split_huge_page* one-liner where needed without
polluting the VM.  Over time we can teach mprotect, mremap and friends to
handle pmd_trans_huge natively without calling split_huge_page*.  The fact
it can't fail isn't just for swap: if split_huge_page would return -ENOMEM
(instead of the current void) we'd need to rollback the mprotect from the
middle of it (ideally including undoing the split_vma) which would be a
big change and in the very wrong direction (it'd likely be simpler not to
call split_huge_page at all and to teach mprotect and friends to handle
hugepages instead of rolling them back from the middle).  In short the
very value of split_huge_page is that it can't fail.

The collapsing and madvise(MADV_HUGEPAGE) part will remain separated and
incremental and it'll just be an "harmless" addition later if this initial
part is agreed upon.  It also should be noted that locking-wise replacing
regular pages with hugepages is going to be very easy if compared to what
I'm doing below in split_huge_page, as it will only happen when
page_count(page) matches page_mapcount(page) if we can take the PG_lock
and mmap_sem in write mode.  collapse_huge_page will be a "best effort"
that (unlike split_huge_page) can fail at the minimal sign of trouble and
we can try again later.  collapse_huge_page will be similar to how KSM
works and the madvise(MADV_HUGEPAGE) will work similar to
madvise(MADV_MERGEABLE).

The default I like is that transparent hugepages are used at page fault
time.  This can be changed with
/sys/kernel/mm/transparent_hugepage/enabled.  The control knob can be set
to three values "always", "madvise", "never" which mean respectively that
hugepages are always used, or only inside madvise(MADV_HUGEPAGE) regions,
or never used.  /sys/kernel/mm/transparent_hugepage/defrag instead
controls if the hugepage allocation should defrag memory aggressively
"always", only inside "madvise" regions, or "never".

The pmd_trans_splitting/pmd_trans_huge locking is very solid.  The
put_page (from get_user_page users that can't use mmu notifier like
O_DIRECT) that runs against a __split_huge_page_refcount instead was a
pain to serialize in a way that would result always in a coherent page
count for both tail and head.  I think my locking solution with a
compound_lock taken only after the page_first is valid and is still a
PageHead should be safe but it surely needs review from SMP race point of
view.  In short there is no current existing way to serialize the O_DIRECT
final put_page against split_huge_page_refcount so I had to invent a new
one (O_DIRECT loses knowledge on the mapping status by the time gup_fast
returns so...).  And I didn't want to impact all gup/gup_fast users for
now, maybe if we change the gup interface substantially we can avoid this
locking, I admit I didn't think too much about it because changing the gup
unpinning interface would be invasive.

If we ignored O_DIRECT we could stick to the existing compound refcounting
code, by simply adding a get_user_pages_fast_flags(foll_flags) where KVM
(and any other mmu notifier user) would call it without FOLL_GET (and if
FOLL_GET isn't set we'd just BUG_ON if nobody registered itself in the
current task mmu notifier list yet).  But O_DIRECT is fundamental for
decent performance of virtualized I/O on fast storage so we can't avoid it
to solve the race of put_page against split_huge_page_refcount to achieve
a complete hugepage feature for KVM.

Swap and oom works fine (well just like with regular pages ;).  MMU
notifier is handled transparently too, with the exception of the young bit
on the pmd, that didn't have a range check but I think KVM will be fine
because the whole point of hugepages is that EPT/NPT will also use a huge
pmd when they notice gup returns pages with PageCompound set, so they
won't care of a range and there's just the pmd young bit to check in that
case.

NOTE: in some cases if the L2 cache is small, this may slowdown and waste
memory during COWs because 4M of memory are accessed in a single fault
instead of 8k (the payoff is that after COW the program can run faster).
So we might want to switch the copy_huge_page (and clear_huge_page too) to
not temporal stores.  I also extensively researched ways to avoid this
cache trashing with a full prefault logic that would cow in 8k/16k/32k/64k
up to 1M (I can send those patches that fully implemented prefault) but I
concluded they're not worth it and they add an huge additional complexity
and they remove all tlb benefits until the full hugepage has been faulted
in, to save a little bit of memory and some cache during app startup, but
they still don't improve substantially the cache-trashing during startup
if the prefault happens in >4k chunks.  One reason is that those 4k pte
entries copied are still mapped on a perfectly cache-colored hugepage, so
the trashing is the worst one can generate in those copies (cow of 4k page
copies aren't so well colored so they trashes less, but again this results
in software running faster after the page fault).  Those prefault patches
allowed things like a pte where post-cow pages were local 4k regular anon
pages and the not-yet-cowed pte entries were pointing in the middle of
some hugepage mapped read-only.  If it doesn't payoff substantially with
todays hardware it will payoff even less in the future with larger l2
caches, and the prefault logic would blot the VM a lot.  If one is
emebdded transparent_hugepage can be disabled during boot with sysfs or
with the boot commandline parameter transparent_hugepage=0 (or
transparent_hugepage=2 to restrict hugepages inside madvise regions) that
will ensure not a single hugepage is allocated at boot time.  It is simple
enough to just disable transparent hugepage globally and let transparent
hugepages be allocated selectively by applications in the MADV_HUGEPAGE
region (both at page fault time, and if enabled with the
collapse_huge_page too through the kernel daemon).

This patch supports only hugepages mapped in the pmd, archs that have
smaller hugepages will not fit in this patch alone.  Also some archs like
power have certain tlb limits that prevents mixing different page size in
the same regions so they will not fit in this framework that requires
"graceful fallback" to basic PAGE_SIZE in case of physical memory
fragmentation.  hugetlbfs remains a perfect fit for those because its
software limits happen to match the hardware limits.  hugetlbfs also
remains a perfect fit for hugepage sizes like 1GByte that cannot be hoped
to be found not fragmented after a certain system uptime and that would be
very expensive to defragment with relocation, so requiring reservation.
hugetlbfs is the "reservation way", the point of transparent hugepages is
not to have any reservation at all and maximizing the use of cache and
hugepages at all times automatically.

Some performance result:

vmx andrea # LD_PRELOAD=/usr/lib64/libhugetlbfs.so HUGETLB_MORECORE=yes HUGETLB_PATH=/mnt/huge/ ./largep
ages3
memset page fault 1566023
memset tlb miss 453854
memset second tlb miss 453321
random access tlb miss 41635
random access second tlb miss 41658
vmx andrea # LD_PRELOAD=/usr/lib64/libhugetlbfs.so HUGETLB_MORECORE=yes HUGETLB_PATH=/mnt/huge/ ./largepages3
memset page fault 1566471
memset tlb miss 453375
memset second tlb miss 453320
random access tlb miss 41636
random access second tlb miss 41637
vmx andrea # ./largepages3
memset page fault 1566642
memset tlb miss 453417
memset second tlb miss 453313
random access tlb miss 41630
random access second tlb miss 41647
vmx andrea # ./largepages3
memset page fault 1566872
memset tlb miss 453418
memset second tlb miss 453315
random access tlb miss 41618
random access second tlb miss 41659
vmx andrea # echo 0 > /proc/sys/vm/transparent_hugepage
vmx andrea # ./largepages3
memset page fault 2182476
memset tlb miss 460305
memset second tlb miss 460179
random access tlb miss 44483
random access second tlb miss 44186
vmx andrea # ./largepages3
memset page fault 2182791
memset tlb miss 460742
memset second tlb miss 459962
random access tlb miss 43981
random access second tlb miss 43988

============
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/time.h>

#define SIZE (3UL*1024*1024*1024)

int main()
{
	char *p = malloc(SIZE), *p2;
	struct timeval before, after;

	gettimeofday(&before, NULL);
	memset(p, 0, SIZE);
	gettimeofday(&after, NULL);
	printf("memset page fault %Lu\n",
	       (after.tv_sec-before.tv_sec)*1000000UL +
	       after.tv_usec-before.tv_usec);

	gettimeofday(&before, NULL);
	memset(p, 0, SIZE);
	gettimeofday(&after, NULL);
	printf("memset tlb miss %Lu\n",
	       (after.tv_sec-before.tv_sec)*1000000UL +
	       after.tv_usec-before.tv_usec);

	gettimeofday(&before, NULL);
	memset(p, 0, SIZE);
	gettimeofday(&after, NULL);
	printf("memset second tlb miss %Lu\n",
	       (after.tv_sec-before.tv_sec)*1000000UL +
	       after.tv_usec-before.tv_usec);

	gettimeofday(&before, NULL);
	for (p2 = p; p2 < p+SIZE; p2 += 4096)
		*p2 = 0;
	gettimeofday(&after, NULL);
	printf("random access tlb miss %Lu\n",
	       (after.tv_sec-before.tv_sec)*1000000UL +
	       after.tv_usec-before.tv_usec);

	gettimeofday(&before, NULL);
	for (p2 = p; p2 < p+SIZE; p2 += 4096)
		*p2 = 0;
	gettimeofday(&after, NULL);
	printf("random access second tlb miss %Lu\n",
	       (after.tv_sec-before.tv_sec)*1000000UL +
	       after.tv_usec-before.tv_usec);

	return 0;
}
============
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Not worth throwing away the precious reserved free memory pool for
allocations that can fail gracefully (either through mempool or because
they're transhuge allocations later falling back to 4k allocations).
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Reviewed-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Transparent hugepage allocations must be allowed not to invoke kswapd or
any other kind of indirect reclaim (especially when the defrag sysfs is
control disabled).  It's unacceptable to swap out anonymous pages
(potentially anonymous transparent hugepages) in order to create new
transparent hugepages.  This is true for the MADV_HUGEPAGE areas too
(swapping out a kvm virtual machine and so having it suffer an unbearable
slowdown, so another one with guest physical memory marked MADV_HUGEPAGE
can run 30% faster if it is running memory intensive workloads, makes no
sense).  If a transparent hugepage allocation fails the slowdown is minor
and there is total fallback, so kswapd should never be asked to swapout
memory to allow the high order allocation to succeed.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NRik van Riel <riel@redhat.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This should work for both hugetlbfs and transparent hugepages.

[akpm@linux-foundation.org: bring forward PageTransCompound() addition for bisectability]
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Cc: Avi Kivity <avi@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>