These days, hugepages are demand-allocated at first fault time.  There's a
somewhat dubious (and racy) heuristic when making a new mmap() to check if
there are enough available hugepages to fully satisfy that mapping.

A particularly obvious case where the heuristic breaks down is where a
process maps its hugepages not as a single chunk, but as a bunch of
individually mmap()ed (or shmat()ed) blocks without touching and
instantiating the pages in between allocations.  In this case the size of
each block is compared against the total number of available hugepages.
It's thus easy for the process to become overcommitted, because each block
mapping will succeed, although the total number of hugepages required by
all blocks exceeds the number available.  In particular, this defeats such
a program which will detect a mapping failure and adjust its hugepage usage
downward accordingly.

The patch below addresses this problem, by strictly reserving a number of
physical hugepages for hugepage inodes which have been mapped, but not
instatiated.  MAP_SHARED mappings are thus "safe" - they will fail on
mmap(), not later with an OOM SIGKILL.  MAP_PRIVATE mappings can still
trigger an OOM.  (Actually SHARED mappings can technically still OOM, but
only if the sysadmin explicitly reduces the hugepage pool between mapping
and instantiation)

This patch appears to address the problem at hand - it allows DB2 to start
correctly, for instance, which previously suffered the failure described
above.

This patch causes no regressions on the libhugetblfs testsuite, and makes a
test (designed to catch this problem) pass which previously failed (ppc64,
POWER5).
Signed-off-by: NDavid Gibson <dwg@au1.ibm.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

2.6.15's hugepage faulting introduced huge_pages_needed accounting into
hugetlbfs: to count how many pages are already in cache, for spot check on
how far a new mapping may be allowed to extend the file.  But it's muddled:
each hugepage found covers HPAGE_SIZE, not PAGE_SIZE.  Once pages were
already in cache, it would overshoot, wrap its hugepages count backwards,
and so fail a harmless repeat mapping with -ENOMEM.  Fixes the problem
found by Don Dupuis.
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Acked-By: NAdam Litke <agl@us.ibm.com>
Acked-by: NWilliam Irwin <wli@holomorphy.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Anything that writes into a tmpfs filesystem is liable to disproportionately
decrease the available memory on a particular node.  Since there's no telling
what sort of application (e.g.  dd/cp/cat) might be dropping large files
there, this lets the admin choose the appropriate default behavior for their
site's situation.

Introduce a tmpfs mount option which allows specifying a memory policy and
a second option to specify the nodelist for that policy.  With the default
policy, tmpfs will behave as it does today.  This patch adds support for
preferred, bind, and interleave policies.

The default policy will cause pages to be added to tmpfs files on the node
which is doing the writing.  Some jobs expect a single process to create
and manage the tmpfs files.  This results in a node which has a
significantly reduced number of free pages.

With this patch, the administrator can specify the policy and nodes for
that policy where they would prefer allocations.

This patch was originally written by Brent Casavant and Hugh Dickins.  I
added support for the bind and preferred policies and the mpol_nodelist
mount option.
Signed-off-by: NBrent Casavant <bcasavan@sgi.com>
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Signed-off-by: NRobin Holt <holt@sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

fs: Use <linux/capability.h> where capable() is used.
Signed-off-by: NRandy Dunlap <rdunlap@xenotime.net>
Acked-by: NTim Schmielau <tim@physik3.uni-rostock.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch converts the inode semaphore to a mutex. I have tested it on
XFS and compiled as much as one can consider on an ia64. Anyway your
luck with it might be different.
Modified-by: NIngo Molnar <mingo@elte.hu>

(finished the conversion)
Signed-off-by: NJes Sorensen <jes@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Implement copy-on-write support for hugetlb mappings so MAP_PRIVATE can be
supported.  This helps us to safely use hugetlb pages in many more
applications.  The patch makes the following changes.  If needed, I also have
it broken out according to the following paragraphs.

1. Add a pair of functions to set/clear write access on huge ptes.  The
   writable check in make_huge_pte is moved out to the caller for use by COW
   later.

2. Hugetlb copy-on-write requires special case handling in the following
   situations:

   - copy_hugetlb_page_range() - Copied pages must be write protected so
     a COW fault will be triggered (if necessary) if those pages are written
     to.

   - find_or_alloc_huge_page() - Only MAP_SHARED pages are added to the
     page cache.  MAP_PRIVATE pages still need to be locked however.

3. Provide hugetlb_cow() and calls from hugetlb_fault() and
   hugetlb_no_page() which handles the COW fault by making the actual copy.

4. Remove the check in hugetlbfs_file_map() so that MAP_PRIVATE mmaps
   will be allowed.  Make MAP_HUGETLB exempt from the depricated VM_RESERVED
   mapping check.
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NAdam Litke <agl@us.ibm.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: "Seth, Rohit" <rohit.seth@intel.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Currently, if a hugetlbfs is mounted without limits (the default), statfs()
will return -1 for max/free/used blocks.  This does not appear to be in
line with normal convention: simple_statfs() and shmem_statfs() both return
0 in similar cases.  Worse, it confuses the translation logic in
put_compat_statfs(), causing it to return -EOVERFLOW on such a mount.

This patch alters hugetlbfs_statfs() to return 0 for max/free/used blocks
on a mount without limits.  Note that we need the test in the patch below,
rather than just using 0 in the sbinfo structure, because the -1 marked in
the free blocks field is used internally to tell the
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch makes a needlessly global function static.
Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Acked-by: NWilliam Irwin <wli@holomorphy.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Basic overcommit checking for hugetlb_file_map() based on an implementation
used with demand faulting in SLES9.

Since demand faulting can't guarantee the availability of pages at mmap
time, this patch implements a basic sanity check to ensure that the number
of huge pages required to satisfy the mmap are currently available.
Despite the obvious race, I think it is a good start on doing proper
accounting.  I'd like to work towards an accounting system that mimics the
semantics of normal pages (especially for the MAP_PRIVATE/COW case).  That
work is underway and builds on what this patch starts.

Huge page shared memory segments are simpler and still maintain their
commit on shmget semantics.
Signed-off-by: NAdam Litke <agl@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Below is a patch to implement demand faulting for huge pages.  The main
motivation for changing from prefaulting to demand faulting is so that huge
page memory areas can be allocated according to NUMA policy.

Thanks to consolidated hugetlb code, switching the behavior requires changing
only one fault handler.  The bulk of the patch just moves the logic from
hugelb_prefault() to hugetlb_pte_fault() and find_get_huge_page().
Signed-off-by: NAdam Litke <agl@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Reformat hugelbfs_forget_inode and add the missing but harmless
write_inode_now call.  It looks the same as generic_forget_inode now except
for the call to truncate_hugepages instead of truncate_inode_pages.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

hugetlbfs_do_delete_inode is the same as generic_delete_inode now, so remove
it in favour of the latter.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Make hugetlbfs looks the same as generic_detelte_inode, fixing a bunch of
missing updates to it at the same time.  Rename it to
hugetlbfs_do_delete_inode and add a real hugetlbfs_delete_inode that
implements ->delete_inode.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Move hugetlbfs accounting into ->alloc_inode / ->destroy_inode.  This keeps
the code simpler, fixes a loeak where a failing inode allocation wouldn't
decrement the counter and moves hugetlbfs_delete_inode and
hugetlbfs_forget_inode closer to their generic counterparts.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Remove the page_table_lock from around the calls to unmap_vmas, and replace
the pte_offset_map in zap_pte_range by pte_offset_map_lock: all callers are
now safe to descend without page_table_lock.

Don't attempt fancy locking for hugepages, just take page_table_lock in
unmap_hugepage_range.  Which makes zap_hugepage_range, and the hugetlb test in
zap_page_range, redundant: unmap_vmas calls unmap_hugepage_range anyway.  Nor
does unmap_vmas have much use for its mm arg now.

The tlb_start_vma and tlb_end_vma in unmap_page_range are now called without
page_table_lock: if they're implemented at all, they typically come down to
flush_cache_range (usually done outside page_table_lock) and flush_tlb_range
(which we already audited for the mprotect case).
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Ingo recently introduced a great speedup for allocating new mmaps using the
free_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and
causes huge performance increases in thread creation.

The downside of this patch is that it does lead to fragmentation in the
mmap-ed areas (visible via /proc/self/maps), such that some applications
that work fine under 2.4 kernels quickly run out of memory on any 2.6
kernel.

The problem is twofold:

  1) the free_area_cache is used to continue a search for memory where
     the last search ended.  Before the change new areas were always
     searched from the base address on.

     So now new small areas are cluttering holes of all sizes
     throughout the whole mmap-able region whereas before small holes
     tended to close holes near the base leaving holes far from the base
     large and available for larger requests.

  2) the free_area_cache also is set to the location of the last
     munmap-ed area so in scenarios where we allocate e.g.  five regions of
     1K each, then free regions 4 2 3 in this order the next request for 1K
     will be placed in the position of the old region 3, whereas before we
     appended it to the still active region 1, placing it at the location
     of the old region 2.  Before we had 1 free region of 2K, now we only
     get two free regions of 1K -> fragmentation.

The patch addresses thes issues by introducing yet another cache descriptor
cached_hole_size that contains the largest known hole size below the
current free_area_cache.  If a new request comes in the size is compared
against the cached_hole_size and if the request can be filled with a hole
below free_area_cache the search is started from the base instead.

The results look promising: Whereas 2.6.12-rc4 fragments quickly and my
(earlier posted) leakme.c test program terminates after 50000+ iterations
with 96 distinct and fragmented maps in /proc/self/maps it performs nicely
(as expected) with thread creation, Ingo's test_str02 with 20000 threads
requires 0.7s system time.

Taking out Ingo's patch (un-patch available per request) by basically
deleting all mentions of free_area_cache from the kernel and starting the
search for new memory always at the respective bases we observe: leakme
terminates successfully with 11 distinctive hardly fragmented areas in
/proc/self/maps but thread creating is gringdingly slow: 30+s(!) system
time for Ingo's test_str02 with 20000 threads.

Now - drumroll ;-) the appended patch works fine with leakme: it ends with
only 7 distinct areas in /proc/self/maps and also thread creation seems
sufficiently fast with 0.71s for 20000 threads.
Signed-off-by: NWolfgang Wander <wwc@rentec.com>
Credit-to: "Richard Purdie" <rpurdie@rpsys.net>
Signed-off-by: NKen Chen <kenneth.w.chen@intel.com>
Acked-by: Ingo Molnar <mingo@elte.hu> (partly)
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!