Commit Graph

377 Commits

Author SHA1 Message Date
David Rientjes
e33c3b5e17 cpusets: update mems allowed in page allocator
If all allowable memory is unreclaimable, it is possible to loop forever
in the page allocator for ~__GFP_NORETRY allocations.

During this time, it is also possible for a task's cpuset to expand its
set of allowable nodes so that it now includes free memory.  The cached
copy of this set, current->mems_allowed, is stale, however, since there
has not been a subsequent call to cpuset_update_task_memory_state().

The cached copy of the set of allowable nodes is now updated in the page
allocator's slow path so the additional memory is available to
get_page_from_freelist().

[akpm@linux-foundation.org: add comment]
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Paul Menage <menage@google.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-11-12 17:17:16 -08:00
Andy Whitcroft
18229df5b6 hugetlb: pull gigantic page initialisation out of the default path
As we can determine exactly when a gigantic page is in use we can optimise
the common regular page cases by pulling out gigantic page initialisation
into its own function.  As gigantic pages are never released to buddy we
do not need a destructor.  This effectivly reverts the previous change to
the main buddy allocator.  It also adds a paranoid check to ensure we
never release gigantic pages from hugetlbfs to the main buddy.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Cc: Jon Tollefson <kniht@linux.vnet.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: <stable@kernel.org>		[2.6.27.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-11-06 15:41:18 -08:00
KAMEZAWA Hiroyuki
52d4b9ac0b memcg: allocate all page_cgroup at boot
Allocate all page_cgroup at boot and remove page_cgroup poitner from
struct page.  This patch adds an interface as

 struct page_cgroup *lookup_page_cgroup(struct page*)

All FLATMEM/DISCONTIGMEM/SPARSEMEM  and MEMORY_HOTPLUG is supported.

Remove page_cgroup pointer reduces the amount of memory by
 - 4 bytes per PAGE_SIZE.
 - 8 bytes per PAGE_SIZE
if memory controller is disabled. (even if configured.)

On usual 8GB x86-32 server, this saves 8MB of NORMAL_ZONE memory.
On my x86-64 server with 48GB of memory, this saves 96MB of memory.
I think this reduction makes sense.

By pre-allocation, kmalloc/kfree in charge/uncharge are removed.
This means
  - we're not necessary to be afraid of kmalloc faiulre.
    (this can happen because of gfp_mask type.)
  - we can avoid calling kmalloc/kfree.
  - we can avoid allocating tons of small objects which can be fragmented.
  - we can know what amount of memory will be used for this extra-lru handling.

I added printk message as

	"allocated %ld bytes of page_cgroup"
        "please try cgroup_disable=memory option if you don't want"

maybe enough informative for users.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:39 -07:00
Gerald Schaefer
1125b4e394 setup_per_zone_pages_min(): take zone->lock instead of zone->lru_lock
This replaces zone->lru_lock in setup_per_zone_pages_min() with zone->lock.
There seems to be no need for the lru_lock anymore, but there is a need for
zone->lock instead, because that function may call move_freepages() via
setup_zone_migrate_reserve().

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Tested-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:32 -07:00
Yinghai Lu
d903ef9f38 mm: print out meminit for memmap
Improve debuggability of memory setup problems.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:32 -07:00
Lee Schermerhorn
985737cf2e mlock: count attempts to free mlocked page
Allow free of mlock()ed pages.  This shouldn't happen, but during
developement, it occasionally did.

This patch allows us to survive that condition, while keeping the
statistics and events correct for debug.

Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:31 -07:00
Nick Piggin
b291f00039 mlock: mlocked pages are unevictable
Make sure that mlocked pages also live on the unevictable LRU, so kswapd
will not scan them over and over again.

This is achieved through various strategies:

1) add yet another page flag--PG_mlocked--to indicate that
   the page is locked for efficient testing in vmscan and,
   optionally, fault path.  This allows early culling of
   unevictable pages, preventing them from getting to
   page_referenced()/try_to_unmap().  Also allows separate
   accounting of mlock'd pages, as Nick's original patch
   did.

   Note:  Nick's original mlock patch used a PG_mlocked
   flag.  I had removed this in favor of the PG_unevictable
   flag + an mlock_count [new page struct member].  I
   restored the PG_mlocked flag to eliminate the new
   count field.

2) add the mlock/unevictable infrastructure to mm/mlock.c,
   with internal APIs in mm/internal.h.  This is a rework
   of Nick's original patch to these files, taking into
   account that mlocked pages are now kept on unevictable
   LRU list.

3) update vmscan.c:page_evictable() to check PageMlocked()
   and, if vma passed in, the vm_flags.  Note that the vma
   will only be passed in for new pages in the fault path;
   and then only if the "cull unevictable pages in fault
   path" patch is included.

4) add try_to_unlock() to rmap.c to walk a page's rmap and
   ClearPageMlocked() if no other vmas have it mlocked.
   Reuses as much of try_to_unmap() as possible.  This
   effectively replaces the use of one of the lru list links
   as an mlock count.  If this mechanism let's pages in mlocked
   vmas leak through w/o PG_mlocked set [I don't know that it
   does], we should catch them later in try_to_unmap().  One
   hopes this will be rare, as it will be relatively expensive.

Original mm/internal.h, mm/rmap.c and mm/mlock.c changes:
Signed-off-by: Nick Piggin <npiggin@suse.de>

splitlru: introduce __get_user_pages():

  New munlock processing need to GUP_FLAGS_IGNORE_VMA_PERMISSIONS.
  because current get_user_pages() can't grab PROT_NONE pages theresore it
  cause PROT_NONE pages can't munlock.

[akpm@linux-foundation.org: fix this for pagemap-pass-mm-into-pagewalkers.patch]
[akpm@linux-foundation.org: untangle patch interdependencies]
[akpm@linux-foundation.org: fix things after out-of-order merging]
[hugh@veritas.com: fix page-flags mess]
[lee.schermerhorn@hp.com: fix munlock page table walk - now requires 'mm']
[kosaki.motohiro@jp.fujitsu.com: build fix]
[kosaki.motohiro@jp.fujitsu.com: fix truncate race and sevaral comments]
[kosaki.motohiro@jp.fujitsu.com: splitlru: introduce __get_user_pages()]
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: Matt Mackall <mpm@selenic.com>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:30 -07:00
Lee Schermerhorn
7b854121eb Unevictable LRU Page Statistics
Report unevictable pages per zone and system wide.

Kosaki Motohiro added support for memory controller unevictable
statistics.

[riel@redhat.com: fix printk in show_free_areas()]
[akpm@linux-foundation.org: fix units in /proc/vmstats]
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Debugged-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:50:26 -07:00
Rik van Riel
556adecba1 vmscan: second chance replacement for anonymous pages
We avoid evicting and scanning anonymous pages for the most part, but
under some workloads we can end up with most of memory filled with
anonymous pages.  At that point, we suddenly need to clear the referenced
bits on all of memory, which can take ages on very large memory systems.

We can reduce the maximum number of pages that need to be scanned by not
taking the referenced state into account when deactivating an anonymous
page.  After all, every anonymous page starts out referenced, so why
check?

If an anonymous page gets referenced again before it reaches the end of
the inactive list, we move it back to the active list.

To keep the maximum amount of necessary work reasonable, we scale the
active to inactive ratio with the size of memory, using the formula
active:inactive ratio = sqrt(memory in GB * 10).

Kswapd CPU use now seems to scale by the amount of pageout bandwidth,
instead of by the amount of memory present in the system.

[kamezawa.hiroyu@jp.fujitsu.com: fix OOM with memcg]
[kamezawa.hiroyu@jp.fujitsu.com: memcg: lru scan fix]
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:50:25 -07:00
Rik van Riel
4f98a2fee8 vmscan: split LRU lists into anon & file sets
Split the LRU lists in two, one set for pages that are backed by real file
systems ("file") and one for pages that are backed by memory and swap
("anon").  The latter includes tmpfs.

The advantage of doing this is that the VM will not have to scan over lots
of anonymous pages (which we generally do not want to swap out), just to
find the page cache pages that it should evict.

This patch has the infrastructure and a basic policy to balance how much
we scan the anon lists and how much we scan the file lists.  The big
policy changes are in separate patches.

[lee.schermerhorn@hp.com: collect lru meminfo statistics from correct offset]
[kosaki.motohiro@jp.fujitsu.com: prevent incorrect oom under split_lru]
[kosaki.motohiro@jp.fujitsu.com: fix pagevec_move_tail() doesn't treat unevictable page]
[hugh@veritas.com: memcg swapbacked pages active]
[hugh@veritas.com: splitlru: BDI_CAP_SWAP_BACKED]
[akpm@linux-foundation.org: fix /proc/vmstat units]
[nishimura@mxp.nes.nec.co.jp: memcg: fix handling of shmem migration]
[kosaki.motohiro@jp.fujitsu.com: adjust Quicklists field of /proc/meminfo]
[kosaki.motohiro@jp.fujitsu.com: fix style issue of get_scan_ratio()]
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:50:25 -07:00
Rik van Riel
b2e185384f define page_file_cache() function
Define page_file_cache() function to answer the question:
	is page backed by a file?

Originally part of Rik van Riel's split-lru patch.  Extracted to make
available for other, independent reclaim patches.

Moved inline function to linux/mm_inline.h where it will be needed by
subsequent "split LRU" and "noreclaim" patches.

Unfortunately this needs to use a page flag, since the PG_swapbacked state
needs to be preserved all the way to the point where the page is last
removed from the LRU.  Trying to derive the status from other info in the
page resulted in wrong VM statistics in earlier split VM patchsets.

The total number of page flags in use on a 32 bit machine after this patch
is 19.

[akpm@linux-foundation.org: fix up out-of-order merge fallout]
[hugh@veritas.com: splitlru: shmem_getpage SetPageSwapBacked sooner[
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: MinChan Kim <minchan.kim@gmail.com>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:50:25 -07:00
Christoph Lameter
b69408e88b vmscan: Use an indexed array for LRU variables
Currently we are defining explicit variables for the inactive and active
list.  An indexed array can be more generic and avoid repeating similar
code in several places in the reclaim code.

We are saving a few bytes in terms of code size:

Before:

   text    data     bss     dec     hex filename
4097753  573120 4092484 8763357  85b7dd vmlinux

After:

   text    data     bss     dec     hex filename
4097729  573120 4092484 8763333  85b7c5 vmlinux

Having an easy way to add new lru lists may ease future work on the
reclaim code.

Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:50:25 -07:00
Andrew Morton
db99100d2e mm/page_alloc.c:free_area_init_nodes() fix inappropriate use of enum
Local variable `i' is a) misleadingly-named for an `enum zone_type' and b)
used for indexing zones as well as nodes as well as node_maps.

Make it an `int'.

Reported-by: Frans Pop <elendil@planet.nl>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-16 11:21:29 -07:00
Andy Whitcroft
6babc32c41 mm: handle initialising compound pages at orders greater than MAX_ORDER
When we initialise a compound page we initialise the page flags and head
page pointer for all base pages spanned by that page.  When we initialise
a gigantic page (a page of order greater than or equal to MAX_ORDER) we
have to initialise more than MAX_ORDER_NR_PAGES pages.  Currently we
assume that all elements of the mem_map in this page are contigious in
memory.  However this is only guarenteed out to MAX_ORDER_NR_PAGES pages,
and with SPARSEMEM enabled they will not be contigious.  This leads us to
walk off the end of the first section and scribble on everything which
follows, BAD.

When we reach a MAX_ORDER_NR_PAGES boundary we much locate the next
section of the mem_map.  As gigantic pages can only be maximally aligned
we know this will occur at exact multiple of MAX_ORDER_NR_PAGES pages from
the start of the page.

This is a bug fix for the gigantic page support in hugetlbfs.

Credit to Mel Gorman for spotting the issue.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Jon Tollefson <kniht@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-02 15:53:13 -07:00
Marcin Slusarz
527655835e mm/bootmem: silence section mismatch warning - contig_page_data/bootmem_node_data
WARNING: vmlinux.o(.data+0x1f5c0): Section mismatch in reference from the variable contig_page_data to the variable .init.data:bootmem_node_data
The variable contig_page_data references
the variable __initdata bootmem_node_data
If the reference is valid then annotate the
variable with __init* (see linux/init.h) or name the variable:
*driver, *_template, *_timer, *_sht, *_ops, *_probe, *_probe_one, *_console,

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Cc: Johannes Weiner <hannes@saeurebad.de>
Cc: Sean MacLennan <smaclennan@pikatech.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-02 19:21:37 -07:00
Adam Litke
344c790e38 mm: make setup_zone_migrate_reserve() aware of overlapping nodes
I have gotten to the root cause of the hugetlb badness I reported back on
August 15th.  My system has the following memory topology (note the
overlapping node):

            Node 0 Memory: 0x8000000-0x44000000
            Node 1 Memory: 0x0-0x8000000 0x44000000-0x80000000

setup_zone_migrate_reserve() scans the address range 0x0-0x8000000 looking
for a pageblock to move onto the MIGRATE_RESERVE list.  Finding no
candidates, it happily continues the scan into 0x8000000-0x44000000.  When
a pageblock is found, the pages are moved to the MIGRATE_RESERVE list on
the wrong zone.  Oops.

setup_zone_migrate_reserve() should skip pageblocks in overlapping nodes.

Signed-off-by: Adam Litke <agl@us.ibm.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: Nishanth Aravamudan <nacc@us.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: <stable@kernel.org>		[2.6.25.x, 2.6.26.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-02 19:21:37 -07:00
Jan Beulich
74768ed833 page allocator: use no-panic variant of alloc_bootmem() in alloc_large_system_hash()
..  since a failed allocation is being (initially) handled gracefully, and
panic()-ed upon failure explicitly in the function if retries with smaller
sizes failed.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-08-12 16:07:27 -07:00
Yinghai Lu
1d1958f050 mm: remove find_max_pfn_with_active_regions
It has no user now

Also print out info about adding/removing active regions.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-30 09:41:44 -07:00
Rusty Russell
9b1a4d3837 stop_machine: Wean existing callers off stop_machine_run()
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2008-07-28 12:16:31 +10:00
Yasunori Goto
af370fb8cb memory hotplug: small fixes to bootmem freeing for memory hotremove
- Change some naming
  * Magic -> types
  * MIX_INFO -> MIX_SECTION_INFO
  * Change definition of bootmem type from direct hex value

- __free_pages_bootmem() becomes __meminit.

Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Johannes Weiner <hannes@saeurebad.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:21 -07:00
Adrian Bunk
b69a7288ea mm/page_alloc.c: cleanups
This patch contains the following cleanups:
- make the following needlessly global variables static:
  - required_kernelcore
  - zone_movable_pfn[]
- make the following needlessly global functions static:
  - move_freepages()
  - move_freepages_block()
  - setup_pageset()
  - find_usable_zone_for_movable()
  - adjust_zone_range_for_zone_movable()
  - __absent_pages_in_range()
  - find_min_pfn_for_node()
  - find_zone_movable_pfns_for_nodes()

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:20 -07:00
Timur Tabi
2be0ffe2b2 mm: add alloc_pages_exact() and free_pages_exact()
alloc_pages_exact() is similar to alloc_pages(), except that it allocates
the minimum number of pages to fulfill the request.  This is useful if you
want to allocate a very large buffer that is slightly larger than an even
power-of-two number of pages.  In that case, alloc_pages() will waste a
lot of memory.

I have a video driver that wants to allocate a 5MB buffer.  alloc_pages()
wiill waste 3MB of physically-contiguous memory.

Signed-off-by: Timur Tabi <timur@freescale.com>
Cc: Andi Kleen <andi@firstfloor.org>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:20 -07:00
Andi Kleen
01ad1c0827 mm: export prep_compound_page to mm
hugetlb will need to get compound pages from bootmem to handle the case of
them being greater than or equal to MAX_ORDER.  Export the constructor
function needed for this.

Acked-by: Adam Litke <agl@us.ibm.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:17 -07:00
Johannes Weiner
9109fb7b35 mm: drop unneeded pgdat argument from free_area_init_node()
free_area_init_node() gets passed in the node id as well as the node
descriptor.  This is redundant as the function can trivially get the node
descriptor itself by means of NODE_DATA() and the node's id.

I checked all the users and NODE_DATA() seems to be usable everywhere
from where this function is called.

Signed-off-by: Johannes Weiner <hannes@saeurebad.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:16 -07:00
Andy Whitcroft
3c82d0ce2c buddy: clarify comments describing buddy merge
In __free_one_page(), the comment "Move the buddy up one level" appears
attached to the break and by implication when the break is taken we are
moving it up one level:

	if (!page_is_buddy(page, buddy, order))
		break;          /* Move the buddy up one level. */

In reality the inverse is true, we break out when we can no longer merge
this page with its buddy.  Looking back into pre-history (into the full
git history) it appears that these two lines accidentally got joined as
part of another change.

Move the comment down where it belongs below the if and clarify its
language.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:15 -07:00
KOSAKI Motohiro
e4048e5dc4 page allocator: inline some __alloc_pages() wrappers
Two zonelist patch series rewrote __page_alloc() largely.  Now, it is just
a wrapper function.  Inlining them will save a function call.

[akpm@linux-foundation.org: export __alloc_pages_internal]
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:14 -07:00
Johannes Weiner
b61bfa3c46 mm: move bootmem descriptors definition to a single place
There are a lot of places that define either a single bootmem descriptor or an
array of them.  Use only one central array with MAX_NUMNODES items instead.

Signed-off-by: Johannes Weiner <hannes@saeurebad.de>
Acked-by: Ralf Baechle <ralf@linux-mips.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Hirokazu Takata <takata@linux-m32r.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Kyle McMartin <kyle@parisc-linux.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:14 -07:00
Mel Gorman
68ad8df42e mm: print out the zonelists on request for manual verification
This patch prints out the zonelists during boot for manual verification by the
user if the mminit_loglevel is MMINIT_VERIFY or higher.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:14 -07:00
Mel Gorman
2dbb51c49f mm: make defensive checks around PFN values registered for memory usage
There are a number of different views to how much memory is currently active.
There is the arch-independent zone-sizing view, the bootmem allocator and
memory models view.

Architectures register this information at different times and is not
necessarily in sync particularly with respect to some SPARSEMEM limitations.

This patch introduces mminit_validate_memmodel_limits() which is able to
validate and correct PFN ranges with respect to the memory model.  It is only
SPARSEMEM that currently validates itself.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:13 -07:00
Mel Gorman
708614e618 mm: verify the page links and memory model
Print out information on how the page flags are being used if mminit_loglevel
is MMINIT_VERIFY or higher and unconditionally performs sanity checks on the
flags regardless of loglevel.

When the page flags are updated with section, node and zone information, a
check are made to ensure the values can be retrieved correctly.  Finally we
confirm that pfn_to_page and page_to_pfn are the correct inverse functions.

[akpm@linux-foundation.org: fix printk warnings]
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:13 -07:00
Mel Gorman
6b74ab97bc mm: add a basic debugging framework for memory initialisation
Boot initialisation is very complex, with significant numbers of
architecture-specific routines, hooks and code ordering.  While significant
amounts of the initialisation is architecture-independent, it trusts the data
received from the architecture layer.  This is a mistake, and has resulted in
a number of difficult-to-diagnose bugs.

This patchset adds some validation and tracing to memory initialisation.  It
also introduces a few basic defensive measures.  The validation code can be
explicitly disabled for embedded systems.

This patch:

Add additional debugging and verification code for memory initialisation.

Once enabled, the verification checks are always run and when required
additional debugging information may be outputted via a mminit_loglevel=
command-line parameter.

The verification code is placed in a new file mm/mm_init.c.  Ideally other mm
initialisation code will be moved here over time.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:13 -07:00
Ingo Molnar
1a781a777b Merge branch 'generic-ipi' into generic-ipi-for-linus
Conflicts:

	arch/powerpc/Kconfig
	arch/s390/kernel/time.c
	arch/x86/kernel/apic_32.c
	arch/x86/kernel/cpu/perfctr-watchdog.c
	arch/x86/kernel/i8259_64.c
	arch/x86/kernel/ldt.c
	arch/x86/kernel/nmi_64.c
	arch/x86/kernel/smpboot.c
	arch/x86/xen/smp.c
	include/asm-x86/hw_irq_32.h
	include/asm-x86/hw_irq_64.h
	include/asm-x86/mach-default/irq_vectors.h
	include/asm-x86/mach-voyager/irq_vectors.h
	include/asm-x86/smp.h
	kernel/Makefile

Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-15 21:55:59 +02:00
Paul Jackson
5dab8ec139 mm, generic, x86 boot: more tweaks to hex prints of some pfn addresses
Fix some problems with (and applies on top of) a previous patch:
  x86 boot: show pfn addresses in hex not decimal in some kernel info printks

Primarily change "0x%8lx" format, which displays with a right aligned
space filled hex number (spaces between the "0x" prefix and the number),
into "%0#10lx" format, which zero fills instead of space fills, and
which uses the printf flag '#' to request the "0x" prefix instead of
hard coding it.

Also replace some other "0x%lx" formats with "%#lx", making use of the
'#' printf flag again.

Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: "Yinghai Lu" <yhlu.kernel@gmail.com>
Cc: "Jack Steiner" <steiner@sgi.com>
Cc: "Mike Travis" <travis@sgi.com>
Cc: "Huang
Cc: Ying" <ying.huang@intel.com>
Cc: "Andi Kleen" <andi@firstfloor.org>
Cc: "Andrew Morton" <akpm@linux-foundation.org>
Cc: Paul Jackson <pj@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-08 13:10:40 +02:00
Paul Jackson
2bc0d2615a x86 boot: more consistently use type int for node ids
Everywhere I look, node id's are of type 'int', except in this one
case, which has 'unsigned long'.  Change this one to 'int' as well.
There is nothing special about the way this variable 'nid' is used in
this routine to justify using an unusual type here.

Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: "Yinghai Lu" <yhlu.kernel@gmail.com>
Cc: "Jack Steiner" <steiner@sgi.com>
Cc: "Mike Travis" <travis@sgi.com>
Cc: "Huang
Cc: Ying" <ying.huang@intel.com>
Cc: "Andi Kleen" <andi@firstfloor.org>
Cc: "Andrew Morton" <akpm@linux-foundation.org>
Cc: Paul Jackson <pj@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-08 12:51:38 +02:00
Paul Jackson
e2fc252e0c x86 boot: show pfn addresses in hex not decimal in some kernel info printks
Page frame numbers (the portion of physical addresses above the low
order page offsets) are displayed in several kernel debug and info
prints in decimal, not hex.  Decimal addresse are unreadable.  Use hex.

Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: "Yinghai Lu" <yhlu.kernel@gmail.com>
Cc: "Jack Steiner" <steiner@sgi.com>
Cc: "Mike Travis" <travis@sgi.com>
Cc: "Huang
Cc: Ying" <ying.huang@intel.com>
Cc: "Andi Kleen" <andi@firstfloor.org>
Cc: "Andrew Morton" <akpm@linux-foundation.org>
Cc: Paul Jackson <pj@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-08 12:51:37 +02:00
Yinghai Lu
d52d53b8a5 RFC x86: try to remove arch_get_ram_range
want to remove arch_get_ram_range, and use early_node_map instead.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-08 12:48:27 +02:00
Yinghai Lu
b5bc6c0e55 x86, mm: use add_highpages_with_active_regions() for high pages init v2
use early_node_map to init high pages, so we can remove page_is_ram() and
page_is_reserved_early() in the big loop with add_one_highpage

also remove page_is_reserved_early(), it is not needed anymore.

v2: fix the build of other platforms

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-08 10:37:25 +02:00
Yinghai Lu
cc1050bafe x86: replace shrink_active_range() with remove_active_range()
in case we have kva before ramdisk on a node, we still need to use
those ranges.

v2: reserve_early kva ram area, in case there are holes in highmem, to avoid
    those area could be treat as free high pages.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-08 10:36:29 +02:00
Ingo Molnar
896395c290 Merge branch 'linus' into tmp.x86.mpparse.new 2008-07-08 10:32:56 +02:00
Mel Gorman
494de90098 Do not overwrite nr_zones on !NUMA when initialising zlcache_ptr
The non-NUMA case of build_zonelist_cache() would initialize the
zlcache_ptr for both node_zonelists[] to NULL.

Which is problematic, since non-NUMA only has a single node_zonelists[]
entry, and trying to zero the non-existent second one just overwrote the
nr_zones field instead.

As kswapd uses this value to determine what reclaim work is necessary,
the result is that kswapd never reclaims.  This causes processes to
stall frequently in low-memory situations as they always direct reclaim.
This patch initialises zlcache_ptr correctly.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Tested-by: Dan Williams <dan.j.williams@intel.com>
[ Simplified patch a bit ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-03 09:22:59 -07:00
Jens Axboe
15c8b6c1aa on_each_cpu(): kill unused 'retry' parameter
It's not even passed on to smp_call_function() anymore, since that
was removed. So kill it.

Acked-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-06-26 11:24:38 +02:00
Yinghai Lu
cc1a9d86ce mm, x86: shrink_active_range() should check all
Now we are using register_e820_active_regions() instead of
add_active_range() directly. So end_pfn could be different between the
value in early_node_map to node_end_pfn.

So we need to make shrink_active_range() smarter.

shrink_active_range() is a generic MM function in mm/page_alloc.c but
it is only used on 32-bit x86. Should we move it back to some file in
arch/x86?

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-06-10 11:31:44 +02:00
Russ Anderson
dfa7e20cc0 mm: Minor clean-up of page flags in mm/page_alloc.c
Minor source code cleanup of page flags in mm/page_alloc.c.
Move the definition of the groups of bits to page-flags.h.

The purpose of this clean up is that the next patch will
conditionally add a page flag to the groups.  Doing that
in a header file is cleaner than adding #ifdefs to the
C code.

Signed-off-by: Russ Anderson <rja@sgi.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-06-09 10:22:24 -07:00
Yinghai Lu
e8c27ac919 x86, numa, 32-bit: print out debug info on all kvas
also fix the print out of node_remap_end_vaddr

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-06-03 13:26:26 +02:00
Heiko Carstens
cd94b9dbfa memory hotplug: fix early allocation handling
Trying to add memory via add_memory() from within an initcall function
results in

bootmem alloc of 163840 bytes failed!
Kernel panic - not syncing: Out of memory

This is caused by zone_wait_table_init() which uses system_state to decide
if it should use the bootmem allocator or not.

When initcalls are handled the system_state is still SYSTEM_BOOTING but
the bootmem allocator doesn't work anymore.  So the allocation will fail.

To fix this use slab_is_available() instead as indicator like we do it
everywhere else.

[akpm@linux-foundation.org: coding-style fix]
Reviewed-by: Andy Whitcroft <apw@shadowen.org>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-05-24 09:56:12 -07:00
Andy Whitcroft
7eb54824b7 zonelists: handle a node zonelist with no applicable entries
When booting 2.6.26-rc3 on a multi-node x86_32 numa system we are seeing
panics when trying node local allocations:

 BUG: unable to handle kernel NULL pointer dereference at 0000034c
 IP: [<c1042507>] get_page_from_freelist+0x4a/0x18e
 *pdpt = 00000000013a7001 *pde = 0000000000000000
 Oops: 0000 [#1] SMP
 Modules linked in:

 Pid: 0, comm: swapper Not tainted (2.6.26-rc3-00003-g5abc28d #82)
 EIP: 0060:[<c1042507>] EFLAGS: 00010282 CPU: 0
 EIP is at get_page_from_freelist+0x4a/0x18e
 EAX: c1371ed8 EBX: 00000000 ECX: 00000000 EDX: 00000000
 ESI: f7801180 EDI: 00000000 EBP: 00000000 ESP: c1371ec0
  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
 Process swapper (pid: 0, ti=c1370000 task=c12f5b40 task.ti=c1370000)
 Stack: 00000000 00000000 00000000 00000000 000612d0 000412d0 00000000 000412d0
        f7801180 f7c0101c f7c01018 c10426e4 f7c01018 00000001 00000044 00000000
        00000001 c12f5b40 00000001 00000010 00000000 000412d0 00000286 000412d0
 Call Trace:
  [<c10426e4>] __alloc_pages_internal+0x99/0x378
  [<c10429ca>] __alloc_pages+0x7/0x9
  [<c105e0e8>] kmem_getpages+0x66/0xef
  [<c105ec55>] cache_grow+0x8f/0x123
  [<c105f117>] ____cache_alloc_node+0xb9/0xe4
  [<c105f427>] kmem_cache_alloc_node+0x92/0xd2
  [<c122118c>] setup_cpu_cache+0xaf/0x177
  [<c105e6ca>] kmem_cache_create+0x2c8/0x353
  [<c13853af>] kmem_cache_init+0x1ce/0x3ad
  [<c13755c5>] start_kernel+0x178/0x1ee

This occurs when we are scanning the zonelists looking for a ZONE_NORMAL
page.  In this system there is only ZONE_DMA and ZONE_NORMAL memory on
node 0, all other nodes are mapped above 4GB physical.  Here is a dump
of the zonelists from this system:

    zonelists pgdat=c1400000
     0: c14006c0:2 f7c006c0:2 f7e006c0:2 c1400360:1 c1400000:0
     1: c14006c0:2 c1400360:1 c1400000:0
    zonelists pgdat=f7c00000
     0: f7c006c0:2 f7e006c0:2 c14006c0:2 c1400360:1 c1400000:0
     1: f7c006c0:2
    zonelists pgdat=f7e00000
     0: f7e006c0:2 c14006c0:2 f7c006c0:2 c1400360:1 c1400000:0
     1: f7e006c0:2

When performing a node local allocation we call get_page_from_freelist()
looking for a page.  It in turn calls first_zones_zonelist() which returns
a preferred_zone.  Where there are no applicable zones this will be NULL.
However we use this unconditionally, leading to this panic.

Where there are no applicable zones there is no possibility of a successful
allocation, so simply fail the allocation.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-05-24 09:56:11 -07:00
Johannes Weiner
f723215419 mm: don't drop a partial page in a zone's memory map size
In a zone's present pages number, account for all pages occupied by the
memory map, including a partial.

Signed-off-by: Johannes Weiner <hannes@saeurebad.de>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-05-24 09:56:07 -07:00
Heiko Carstens
76cdd58e55 memory_hotplug: always initialize pageblock bitmap
Trying to online a new memory section that was added via memory hotplug
sometimes results in crashes when the new pages are added via __free_page.
 Reason for that is that the pageblock bitmap isn't initialized and hence
contains random stuff.  That means that get_pageblock_migratetype()
returns also random stuff and therefore

	list_add(&page->lru,
		&zone->free_area[order].free_list[migratetype]);

in __free_one_page() tries to do a list_add to something that isn't even
necessarily a list.

This happens since 86051ca5ea ("mm: fix
usemap initialization") which makes sure that the pageblock bitmap gets
only initialized for pages present in a zone.  Unfortunately for hot-added
memory the zones "grow" after the memmap and the pageblock memmap have
been initialized.  Which means that the new pages have an unitialized
bitmap.  To solve this the calls to grow_zone_span() and grow_pgdat_span()
are moved to __add_zone() just before the initialization happens.

The patch also moves the two functions since __add_zone() is the only
caller and I didn't want to add a forward declaration.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-05-14 19:11:15 -07:00
Thomas Gleixner
3ac7fe5a4a infrastructure to debug (dynamic) objects
We can see an ever repeating problem pattern with objects of any kind in the
kernel:

1) freeing of active objects
2) reinitialization of active objects

Both problems can be hard to debug because the crash happens at a point where
we have no chance to decode the root cause anymore.  One problem spot are
kernel timers, where the detection of the problem often happens in interrupt
context and usually causes the machine to panic.

While working on a timer related bug report I had to hack specialized code
into the timer subsystem to get a reasonable hint for the root cause.  This
debug hack was fine for temporary use, but far from a mergeable solution due
to the intrusiveness into the timer code.

The code further lacked the ability to detect and report the root cause
instantly and keep the system operational.

Keeping the system operational is important to get hold of the debug
information without special debugging aids like serial consoles and special
knowledge of the bug reporter.

The problems described above are not restricted to timers, but timers tend to
expose it usually in a full system crash.  Other objects are less explosive,
but the symptoms caused by such mistakes can be even harder to debug.

Instead of creating specialized debugging code for the timer subsystem a
generic infrastructure is created which allows developers to verify their code
and provides an easy to enable debug facility for users in case of trouble.

The debugobjects core code keeps track of operations on static and dynamic
objects by inserting them into a hashed list and sanity checking them on
object operations and provides additional checks whenever kernel memory is
freed.

The tracked object operations are:
- initializing an object
- adding an object to a subsystem list
- deleting an object from a subsystem list

Each operation is sanity checked before the operation is executed and the
subsystem specific code can provide a fixup function which allows to prevent
the damage of the operation.  When the sanity check triggers a warning message
and a stack trace is printed.

The list of operations can be extended if the need arises.  For now it's
limited to the requirements of the first user (timers).

The core code enqueues the objects into hash buckets.  The hash index is
generated from the address of the object to simplify the lookup for the check
on kfree/vfree.  Each bucket has it's own spinlock to avoid contention on a
global lock.

The debug code can be compiled in without being active.  The runtime overhead
is minimal and could be optimized by asm alternatives.  A kernel command line
option enables the debugging code.

Thanks to Ingo Molnar for review, suggestions and cleanup patches.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Greg KH <greg@kroah.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 08:29:53 -07:00
Nishanth Aravamudan
a41f24ea9f page allocator: smarter retry of costly-order allocations
Because of page order checks in __alloc_pages(), hugepage (and similarly
large order) allocations will not retry unless explicitly marked
__GFP_REPEAT. However, the current retry logic is nearly an infinite
loop (or until reclaim does no progress whatsoever). For these costly
allocations, that seems like overkill and could potentially never
terminate. Mel observed that allowing current __GFP_REPEAT semantics for
hugepage allocations essentially killed the system. I believe this is
because we may continue to reclaim small orders of pages all over, but
never have enough to satisfy the hugepage allocation request. This is
clearly only a problem for large order allocations, of which hugepages
are the most obvious (to me).

Modify try_to_free_pages() to indicate how many pages were reclaimed.
Use that information in __alloc_pages() to eventually fail a large
__GFP_REPEAT allocation when we've reclaimed an order of pages equal to
or greater than the allocation's order. This relies on lumpy reclaim
functioning as advertised. Due to fragmentation, lumpy reclaim may not
be able to free up the order needed in one invocation, so multiple
iterations may be requred. In other words, the more fragmented memory
is, the more retry attempts __GFP_REPEAT will make (particularly for
higher order allocations).

This changes the semantics of __GFP_REPEAT subtly, but *only* for
allocations > PAGE_ALLOC_COSTLY_ORDER. With this patch, for those size
allocations, we will try up to some point (at least 1<<order reclaimed
pages), rather than forever (which is the case for allocations <=
PAGE_ALLOC_COSTLY_ORDER).

This change improves the /proc/sys/vm/nr_hugepages interface with a
follow-on patch that makes pool allocations use __GFP_REPEAT. Rather
than administrators repeatedly echo'ing a particular value into the
sysctl, and forcing reclaim into action manually, this change allows for
the sysctl to attempt a reasonable effort itself. Similarly, dynamic
pool growth should be more successful under load, as lumpy reclaim can
try to free up pages, rather than failing right away.

Choosing to reclaim only up to the order of the requested allocation
strikes a balance between not failing hugepage allocations and returning
to the caller when it's unlikely to every succeed. Because of lumpy
reclaim, if we have freed the order requested, hopefully it has been in
big chunks and those chunks will allow our allocation to succeed. If
that isn't the case after freeing up the current order, I don't think it
is likely to succeed in the future, although it is possible given a
particular fragmentation pattern.

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Tested-by: Mel Gorman <mel@csn.ul.ie>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 08:05:58 -07:00