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>
This patch makes page_cgroup->flags to be atomic_ops and define functions
(and macros) to access it.
Before trying to modify memory resource controller, this atomic operation
on flags is necessary. Most of flags in this patch is for LRU and modfied
under mz->lru_lock but we'll add another flags which is not for LRU soon.
For example, we'll place LOCK bit on flags field. We need atomic
operation to modify LRU bit without LOCK.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-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>
Some obvious optimization to memcg.
I found mem_cgroup_charge_statistics() is a little big (in object) and
does unnecessary address calclation. This patch is for optimization to
reduce the size of this function.
And res_counter_charge() is 'likely' to succeed.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-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>
There are not-on-LRU pages which can be mapped and they are not worth to
be accounted. (becasue we can't shrink them and need dirty codes to
handle specical case) We'd like to make use of usual objrmap/radix-tree's
protcol and don't want to account out-of-vm's control pages.
When special_mapping_fault() is called, page->mapping is tend to be NULL
and it's charged as Anonymous page. insert_page() also handles some
special pages from drivers.
This patch is for avoiding to account special pages.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch tries to make page->mapping to be NULL before
mem_cgroup_uncharge_cache_page() is called.
"page->mapping == NULL" is a good check for "whether the page is still
radix-tree or not". This patch also adds BUG_ON() to
mem_cgroup_uncharge_cache_page();
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
While page-cache's charge/uncharge is done under page_lock(), swap-cache
isn't. (anonymous page is charged when it's newly allocated.)
This patch moves do_swap_page()'s charge() call under lock. I don't see
any bad problem *now* but this fix will be good for future for avoiding
unnecessary racy state.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
do_pages_stat() does not need any page_to_node entry for real. Just pass
the pointers to the user-space page address array and to the user-space
status array, and have do_pages_stat() traverse the former and fill the
latter directly.
Signed-off-by: Brice Goglin <Brice.Goglin@inria.fr>
Acked-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
A patchset reworking sys_move_pages(). It removes the possibly large
vmalloc by using multiple chunks when migrating large buffers. It also
dramatically increases the throughput for large buffers since the lookup
in new_page_node() is now limited to a single chunk, causing the quadratic
complexity to have a much slower impact. There is no need to use any
radix-tree-like structure to improve this lookup.
sys_move_pages() duration on a 4-quadcore-opteron 2347HE (1.9Gz),
migrating between nodes #2 and #3:
length move_pages (us) move_pages+patch (us)
4kB 126 98
40kB 198 168
400kB 963 937
4MB 12503 11930
40MB 246867 11848
Patches #1 and #4 are the important ones:
1) stop returning -ENOENT from sys_move_pages() if nothing got migrated
2) don't vmalloc a huge page_to_node array for do_pages_stat()
3) extract do_pages_move() out of sys_move_pages()
4) rework do_pages_move() to work on page_sized chunks
5) move_pages: no need to set pp->page to ZERO_PAGE(0) by default
This patch:
There is no point in returning -ENOENT from sys_move_pages() if all pages
were already on the right node, while we return 0 if only 1 page was not.
Most application don't know where their pages are allocated, so it's not
an error to try to migrate them anyway.
Just return 0 and let the status array in user-space be checked if the
application needs details.
It will make the upcoming chunked-move_pages() support much easier.
Signed-off-by: Brice Goglin <Brice.Goglin@inria.fr>
Acked-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
During hotplug memory remove, memory regions should be released on a
PAGES_PER_SECTION size chunks. This mirrors the code in add_memory where
resources are requested on a PAGES_PER_SECTION size.
Attempting to release the entire memory region fails because there is not
a single resource for the total number of pages being removed. Instead
the resources for the pages are split in PAGES_PER_SECTION size chunks as
requested during memory add.
Signed-off-by: Nathan Fontenot <nfont@austin.ibm.com>
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Acked-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>
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>
Presently hugepage doesn't use zero page at all because zero page is only
used for coredumping and hugepage can't core dump.
However we have now implemented hugepage coredumping. Therefore we should
implement the zero page of hugepage.
Implementation note:
o Why do we only check VM_SHARED for zero page?
normal page checked as ..
static inline int use_zero_page(struct vm_area_struct *vma)
{
if (vma->vm_flags & (VM_LOCKED | VM_SHARED))
return 0;
return !vma->vm_ops || !vma->vm_ops->fault;
}
First, hugepages are never mlock()ed. We aren't concerned with VM_LOCKED.
Second, hugetlbfs is a pseudo filesystem, not a real filesystem and it
doesn't have any file backing. Thus ops->fault checking is meaningless.
o Why don't we use zero page if !pte.
!pte indicate {pud, pmd} doesn't exist or some error happened. So we
shouldn't return zero page if any error occurred.
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Adam Litke <agl@us.ibm.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Kawai Hidehiro <hidehiro.kawai.ez@hitachi.com>
Cc: Mel Gorman <mel@skynet.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/hugetlb.c:265:17: warning: symbol 'resv_map_alloc' was not declared. Should it be static?
mm/hugetlb.c:277:6: warning: symbol 'resv_map_release' was not declared. Should it be static?
mm/hugetlb.c:292:9: warning: Using plain integer as NULL pointer
mm/hugetlb.c:1750:5: warning: symbol 'unmap_ref_private' was not declared. Should it be static?
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Rewrite the vmap allocator to use rbtrees and lazy tlb flushing, and
provide a fast, scalable percpu frontend for small vmaps (requires a
slightly different API, though).
The biggest problem with vmap is actually vunmap. Presently this requires
a global kernel TLB flush, which on most architectures is a broadcast IPI
to all CPUs to flush the cache. This is all done under a global lock. As
the number of CPUs increases, so will the number of vunmaps a scaled
workload will want to perform, and so will the cost of a global TLB flush.
This gives terrible quadratic scalability characteristics.
Another problem is that the entire vmap subsystem works under a single
lock. It is a rwlock, but it is actually taken for write in all the fast
paths, and the read locking would likely never be run concurrently anyway,
so it's just pointless.
This is a rewrite of vmap subsystem to solve those problems. The existing
vmalloc API is implemented on top of the rewritten subsystem.
The TLB flushing problem is solved by using lazy TLB unmapping. vmap
addresses do not have to be flushed immediately when they are vunmapped,
because the kernel will not reuse them again (would be a use-after-free)
until they are reallocated. So the addresses aren't allocated again until
a subsequent TLB flush. A single TLB flush then can flush multiple
vunmaps from each CPU.
XEN and PAT and such do not like deferred TLB flushing because they can't
always handle multiple aliasing virtual addresses to a physical address.
They now call vm_unmap_aliases() in order to flush any deferred mappings.
That call is very expensive (well, actually not a lot more expensive than
a single vunmap under the old scheme), however it should be OK if not
called too often.
The virtual memory extent information is stored in an rbtree rather than a
linked list to improve the algorithmic scalability.
There is a per-CPU allocator for small vmaps, which amortizes or avoids
global locking.
To use the per-CPU interface, the vm_map_ram / vm_unmap_ram interfaces
must be used in place of vmap and vunmap. Vmalloc does not use these
interfaces at the moment, so it will not be quite so scalable (although it
will use lazy TLB flushing).
As a quick test of performance, I ran a test that loops in the kernel,
linearly mapping then touching then unmapping 4 pages. Different numbers
of tests were run in parallel on an 4 core, 2 socket opteron. Results are
in nanoseconds per map+touch+unmap.
threads vanilla vmap rewrite
1 14700 2900
2 33600 3000
4 49500 2800
8 70631 2900
So with a 8 cores, the rewritten version is already 25x faster.
In a slightly more realistic test (although with an older and less
scalable version of the patch), I ripped the not-very-good vunmap batching
code out of XFS, and implemented the large buffer mapping with vm_map_ram
and vm_unmap_ram... along with a couple of other tricks, I was able to
speed up a large directory workload by 20x on a 64 CPU system. I believe
vmap/vunmap is actually sped up a lot more than 20x on such a system, but
I'm running into other locks now. vmap is pretty well blown off the
profiles.
Before:
1352059 total 0.1401
798784 _write_lock 8320.6667 <- vmlist_lock
529313 default_idle 1181.5022
15242 smp_call_function 15.8771 <- vmap tlb flushing
2472 __get_vm_area_node 1.9312 <- vmap
1762 remove_vm_area 4.5885 <- vunmap
316 map_vm_area 0.2297 <- vmap
312 kfree 0.1950
300 _spin_lock 3.1250
252 sn_send_IPI_phys 0.4375 <- tlb flushing
238 vmap 0.8264 <- vmap
216 find_lock_page 0.5192
196 find_next_bit 0.3603
136 sn2_send_IPI 0.2024
130 pio_phys_write_mmr 2.0312
118 unmap_kernel_range 0.1229
After:
78406 total 0.0081
40053 default_idle 89.4040
33576 ia64_spinlock_contention 349.7500
1650 _spin_lock 17.1875
319 __reg_op 0.5538
281 _atomic_dec_and_lock 1.0977
153 mutex_unlock 1.5938
123 iget_locked 0.1671
117 xfs_dir_lookup 0.1662
117 dput 0.1406
114 xfs_iget_core 0.0268
92 xfs_da_hashname 0.1917
75 d_alloc 0.0670
68 vmap_page_range 0.0462 <- vmap
58 kmem_cache_alloc 0.0604
57 memset 0.0540
52 rb_next 0.1625
50 __copy_user 0.0208
49 bitmap_find_free_region 0.2188 <- vmap
46 ia64_sn_udelay 0.1106
45 find_inode_fast 0.1406
42 memcmp 0.2188
42 finish_task_switch 0.1094
42 __d_lookup 0.0410
40 radix_tree_lookup_slot 0.1250
37 _spin_unlock_irqrestore 0.3854
36 xfs_bmapi 0.0050
36 kmem_cache_free 0.0256
35 xfs_vn_getattr 0.0322
34 radix_tree_lookup 0.1062
33 __link_path_walk 0.0035
31 xfs_da_do_buf 0.0091
30 _xfs_buf_find 0.0204
28 find_get_page 0.0875
27 xfs_iread 0.0241
27 __strncpy_from_user 0.2812
26 _xfs_buf_initialize 0.0406
24 _xfs_buf_lookup_pages 0.0179
24 vunmap_page_range 0.0250 <- vunmap
23 find_lock_page 0.0799
22 vm_map_ram 0.0087 <- vmap
20 kfree 0.0125
19 put_page 0.0330
18 __kmalloc 0.0176
17 xfs_da_node_lookup_int 0.0086
17 _read_lock 0.0885
17 page_waitqueue 0.0664
vmap has gone from being the top 5 on the profiles and flushing the crap
out of all TLBs, to using less than 1% of kernel time.
[akpm@linux-foundation.org: cleanups, section fix]
[akpm@linux-foundation.org: fix build on alpha]
Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Krzysztof Helt <krzysztof.h1@poczta.fm>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
__vma_link_file and expand_downwards functions are not small, yeat they
are marked inline. They probably had one callsite sometime in the past,
but now they have more. In order to prevent similar thing, I also
deinlined expand_upwards, despite it having only pne callsite. Nowadays
gcc auto-inlines such static functions anyway. In find_extend_vma, I
removed one extra level of indirection.
Patch is deliberately generated with -U $BIGNUM to make
it easier to see that functions are big.
Result:
# size */*/mmap.o */vmlinux
text data bss dec hex filename
9514 188 16 9718 25f6 0.org/mm/mmap.o
9237 188 16 9441 24e1 deinline/mm/mmap.o
6124402 858996 389480 7372878 70804e 0.org/vmlinux
6124113 858996 389480 7372589 707f2d deinline/vmlinux
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
trylock_page, unlock_page open and close a critical section. Hence,
we can use the lock bitops to get the desired memory ordering.
Also, mark trylock as likely to succeed (and remove the annotation from
callers).
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>
Setting and clearing the page locked when inserting it into swapcache /
pagecache when it has no other references can use non-atomic page flags
operations because no other CPU may be operating on it at this time.
This saves one atomic operation when inserting a page into pagecache.
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>
Rework Posix error return for mlock().
Posix requires error code for mlock*() system calls for some conditions
that differ from what kernel low level functions, such as
get_user_pages(), return for those conditions. For more info, see:
http://marc.info/?l=linux-kernel&m=121750892930775&w=2
This patch provides the same translation of get_user_pages()
error codes to posix specified error codes in the context
of the mlock rework for unevictable lru.
[akpm@linux-foundation.org: fix build]
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This change is intended to make mlock() error returns correct.
make_page_present() is a lower level function used by more than mlock().
Subsequent patch[es] will add this error return fixup in an mlock specific
path.
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
During each reclaim scan we accumulate scan pressure on unrelated lists
which will result in bogus scans and unwanted reclaims eventually.
Scanning lists with few reclaim candidates results in a lot of rotation
and therefor also disturbs the list balancing, putting even more
pressure on the wrong lists.
In a test-case with much streaming IO, and therefor a crowded inactive
file page list, swapping started because
a) anon pages were reclaimed after swap_cluster_max reclaim
invocations -- nr_scan of this list has just accumulated
b) active file pages were scanned because *their* nr_scan has also
accumulated through the same logic. And this in return created a
lot of rotation for file pages and resulted in a decrease of file
list priority, again increasing the pressure on anon pages.
The result was an evicted working set of anon pages while there were
tons of inactive file pages that should have been taken instead.
Signed-off-by: Johannes Weiner <hannes@saeurebad.de>
Reviewed-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>
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>
This patch adds a function to scan individual or all zones' unevictable
lists and move any pages that have become evictable onto the respective
zone's inactive list, where shrink_inactive_list() will deal with them.
Adds sysctl to scan all nodes, and per node attributes to individual
nodes' zones.
Kosaki: If evictable page found in unevictable lru when write
/proc/sys/vm/scan_unevictable_pages, print filename and file offset of
these pages.
[akpm@linux-foundation.org: fix one CONFIG_MMU=n build error]
[kosaki.motohiro@jp.fujitsu.com: adapt vmscan-unevictable-lru-scan-sysctl.patch to new sysfs API]
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>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.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>
Adrian Bunk
KAMEZAWA Hiroyuki
Brice Goglin
Nathan Fontenot
Gerald Schaefer
KOSAKI Motohiro
Yinghai Lu
Harvey Harrison
Nick Piggin
Denys Vlasenko
Lee Schermerhorn
Johannes Weiner