After making dirty a 100M file, the normal behavior is to start the writeback
for all data after 30s delays. But sometimes the following happens instead:
- after 30s: ~4M
- after 5s: ~4M
- after 5s: all remaining 92M
Some analyze shows that the internal io dispatch queues goes like this:
s_io s_more_io
-------------------------
1) 100M,1K 0
2) 1K 96M
3) 0 96M
1) initial state with a 100M file and a 1K file
2) 4M written, nr_to_write <= 0, so write more
3) 1K written, nr_to_write > 0, no more writes(BUG)
nr_to_write > 0 in (3) fools the upper layer to think that data have all been
written out. The big dirty file is actually still sitting in s_more_io. We
cannot simply splice s_more_io back to s_io as soon as s_io becomes empty, and
let the loop in generic_sync_sb_inodes() continue: this may starve newly
expired inodes in s_dirty. It is also not an option to draw inodes from both
s_more_io and s_dirty, an let the loop go on: this might lead to live locks,
and might also starve other superblocks in sync time(well kupdate may still
starve some superblocks, that's another bug).
We have to return when a full scan of s_io completes. So nr_to_write > 0 does
not necessarily mean that "all data are written". This patch introduces a
flag writeback_control.more_io to indicate this situation. With it the big
dirty file no longer has to wait for the next kupdate invocation 5s later.
Cc: David Chinner <dgc@sgi.com>
Cc: Ken Chen <kenchen@google.com>
Signed-off-by: Fengguang Wu <wfg@mail.ustc.edu.cn>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since nothing earlier than gcc-3.2 is supported for kernel
compilation, that 2.95 hack can be removed.
Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm.h doesn't use directly anything from mutex.h and backing-dev.h, so
remove them and add them back to files which need them.
Cross-compile tested on many configs and archs.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
These aren't modular, so SLAB_PANIC is OK.
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is a writeback-internal marker but we're propagating it all the way back
to userspace!.
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
zone->lock is quite an "inner" lock and mostly constrained to page alloc as
well, so like slab locks, it probably isn't something that is critically
important to document here. However unlike slab locks, zone lock could be
used more widely in future, and page_alloc.c might possibly have more
business to do tricky things with pagecache than does slab. So... I don't
think it hurts to document it.
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>
Introduces new zone flag interface for testing and setting flags:
int zone_test_and_set_flag(struct zone *zone, zone_flags_t flag)
Instead of setting and clearing ZONE_RECLAIM_LOCKED each time shrink_zone() is
called, this flag is test and set before starting zone reclaim. Zone reclaim
starts in __alloc_pages() when a zone's watermark fails and the system is in
zone_reclaim_mode. If it's already in reclaim, there's no need to start again
so it is simply considered full for that allocation attempt.
There is a change of behavior with regard to concurrent zone shrinking. It is
now possible for try_to_free_pages() or kswapd to already be shrinking a
particular zone when __alloc_pages() starts zone reclaim. In this case, it is
possible for two concurrent threads to invoke shrink_zone() for a single zone.
This change forbids a zone to be in zone reclaim twice, which was always the
behavior, but allows for concurrent try_to_free_pages() or kswapd shrinking
when starting zone reclaim.
Cc: Andrea Arcangeli <andrea@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There's no reason to sleep in try_set_zone_oom() or clear_zonelist_oom() if
the lock can't be acquired; it will be available soon enough once the zonelist
scanning is done. All other threads waiting for the OOM killer are also
contingent on the exiting task being able to acquire the lock in
clear_zonelist_oom() so it doesn't make sense to put it to sleep.
Cc: Andrea Arcangeli <andrea@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since no task descriptor's 'cpuset' field is dereferenced in the execution of
the OOM killer anymore, it is no longer necessary to take callback_mutex.
[akpm@linux-foundation.org: restore cpuset_lock for other patches]
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Instead of testing for overlap in the memory nodes of the the nearest
exclusive ancestor of both current and the candidate task, it is better to
simply test for intersection between the task's mems_allowed in their task
descriptors. This does not require taking callback_mutex since it is only
used as a hint in the badness scoring.
Tasks that do not have an intersection in their mems_allowed with the current
task are not explicitly restricted from being OOM killed because it is quite
possible that the candidate task has allocated memory there before and has
since changed its mems_allowed.
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Suppresses the extraneous stack and memory dump when a parallel OOM killing
has been found. There's no need to fill the ring buffer with this information
if its already been printed and the condition that triggered the previous OOM
killer has not yet been alleviated.
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Adds a new sysctl, 'oom_kill_allocating_task', which will automatically kill
the OOM-triggering task instead of scanning through the tasklist to find a
memory-hogging target. This is helpful for systems with an insanely large
number of tasks where scanning the tasklist significantly degrades
performance.
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
A final allocation attempt with a very high watermark needs to be attempted
before invoking out_of_memory(). OOM killer serialization needs to occur
before this final attempt, otherwise tasks attempting to OOM-lock all zones in
its zonelist may spin and acquire the lock unnecessarily after the OOM
condition has already been alleviated.
If the final allocation does succeed, the zonelist is simply OOM-unlocked and
__alloc_pages() returns the page. Otherwise, the OOM killer is invoked.
If the task cannot acquire OOM-locks on all zones in its zonelist, it is put
to sleep and the allocation is retried when it gets rescheduled. One of its
zones is already marked as being in the OOM killer so it'll hopefully be
getting some free memory soon, at least enough to satisfy a high watermark
allocation attempt. This prevents needlessly killing a task when the OOM
condition would have already been alleviated if it had simply been given
enough time.
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
OOM killer synchronization should be done with zone granularity so that memory
policy and cpuset allocations may have their corresponding zones locked and
allow parallel kills for other OOM conditions that may exist elsewhere in the
system. DMA allocations can be targeted at the zone level, which would not be
possible if locking was done in nodes or globally.
Synchronization shall be done with a variation of "trylocks." The goal is to
put the current task to sleep and restart the failed allocation attempt later
if the trylock fails. Otherwise, the OOM killer is invoked.
Each zone in the zonelist that __alloc_pages() was called with is checked for
the newly-introduced ZONE_OOM_LOCKED flag. If any zone has this flag present,
the "trylock" to serialize the OOM killer fails and returns zero. Otherwise,
all the zones have ZONE_OOM_LOCKED set and the try_set_zone_oom() function
returns non-zero.
Cc: Andrea Arcangeli <andrea@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Convert the int all_unreclaimable member of struct zone to unsigned long
flags. This can now be used to specify several different zone flags such as
all_unreclaimable and reclaim_in_progress, which can now be removed and
converted to a per-zone flag.
Flags are set and cleared as follows:
zone_set_flag(struct zone *zone, zone_flags_t flag)
zone_clear_flag(struct zone *zone, zone_flags_t flag)
Defines the first zone flags, ZONE_ALL_UNRECLAIMABLE and ZONE_RECLAIM_LOCKED,
which have the same semantics as the old zone->all_unreclaimable and
zone->reclaim_in_progress, respectively. Also converts all current users that
set or clear either flag to use the new interface.
Helper functions are defined to test the flags:
int zone_is_all_unreclaimable(const struct zone *zone)
int zone_is_reclaim_locked(const struct zone *zone)
All flag operators are of the atomic variety because there are currently
readers that are implemented that do not take zone->lock.
[akpm@linux-foundation.org: add needed include]
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The OOM killer's CONSTRAINT definitions are really more appropriate in an
enum, so define them in include/linux/oom.h.
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Move the OOM killer's extern function prototypes to include/linux/oom.h and
include it where necessary.
[clg@fr.ibm.com: build fix]
Cc: Andrea Arcangeli <andrea@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Cedric Le Goater <clg@fr.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Slab constructors currently have a flags parameter that is never used. And
the order of the arguments is opposite to other slab functions. The object
pointer is placed before the kmem_cache pointer.
Convert
ctor(void *object, struct kmem_cache *s, unsigned long flags)
to
ctor(struct kmem_cache *s, void *object)
throughout the kernel
[akpm@linux-foundation.org: coupla fixes]
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Move irq handling out of new slab into __slab_alloc. That is useful for
Mathieu's cmpxchg_local patchset and also allows us to remove the crude
local_irq_off in early_kmem_cache_alloc().
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Based on ideas of Andrew:
http://marc.info/?l=linux-kernel&m=102912915020543&w=2
Scale the bdi dirty limit inversly with the tasks dirty rate.
This makes heavy writers have a lower dirty limit than the occasional writer.
Andrea proposed something similar:
http://lwn.net/Articles/152277/
The main disadvantage to his patch is that he uses an unrelated quantity to
measure time, which leaves him with a workload dependant tunable. Other than
that the two approaches appear quite similar.
[akpm@linux-foundation.org: fix warning]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Scale writeback cache per backing device, proportional to its writeout speed.
By decoupling the BDI dirty thresholds a number of problems we currently have
will go away, namely:
- mutual interference starvation (for any number of BDIs);
- deadlocks with stacked BDIs (loop, FUSE and local NFS mounts).
It might be that all dirty pages are for a single BDI while other BDIs are
idling. By giving each BDI a 'fair' share of the dirty limit, each one can have
dirty pages outstanding and make progress.
A global threshold also creates a deadlock for stacked BDIs; when A writes to
B, and A generates enough dirty pages to get throttled, B will never start
writeback until the dirty pages go away. Again, by giving each BDI its own
'independent' dirty limit, this problem is avoided.
So the problem is to determine how to distribute the total dirty limit across
the BDIs fairly and efficiently. A DBI that has a large dirty limit but does
not have any dirty pages outstanding is a waste.
What is done is to keep a floating proportion between the DBIs based on
writeback completions. This way faster/more active devices get a larger share
than slower/idle devices.
[akpm@linux-foundation.org: fix warnings]
[hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
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>
Provide scalable per backing_dev_info statistics counters.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
These patches aim to improve balance_dirty_pages() and directly address three
issues:
1) inter device starvation
2) stacked device deadlocks
3) inter process starvation
1 and 2 are a direct result from removing the global dirty limit and using
per device dirty limits. By giving each device its own dirty limit is will
no longer starve another device, and the cyclic dependancy on the dirty limit
is broken.
In order to efficiently distribute the dirty limit across the independant
devices a floating proportion is used, this will allocate a share of the total
limit proportional to the device's recent activity.
3 is done by also scaling the dirty limit proportional to the current task's
recent dirty rate.
This patch:
nfs: remove congestion_end(). It's redundant, clear_bdi_congested() already
wakes the waiters.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch contains the following cleanups:
- make the needlessly global setup_vmstat() static
- remove the unused refresh_vm_stats()
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch contains the following cleanups:
- every file should include the headers containing the prototypes for
its global functions
- make the follosing needlessly global functions static:
- migrate_to_node()
- do_mbind()
- sp_alloc()
- mpol_rebind_policy()
[akpm@linux-foundation.org: fix uninitialised var warning]
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch makes three needlessly global functions static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When gather_surplus_pages() fails to allocate enough huge pages to satisfy
the requested reservation, it frees what it did allocate back to the buddy
allocator. put_page() should be called instead of update_and_free_page()
to ensure that pool counters are updated as appropriate and the page's
refcount is decremented.
Signed-off-by: Adam Litke <agl@us.ibm.com>
Acked-by: Dave Hansen <haveblue@us.ibm.com>
Cc: David Gibson <hermes@gibson.dropbear.id.au>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Ken Chen <kenchen@google.com>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton found a problem with the hugetlb pool allocation when some nodes have
no memory (http://marc.info/?l=linux-mm&m=118133042025995&w=2). Lee worked
on versions that tried to fix it, but none were accepted. Christoph has
created a set of patches which allow for GFP_THISNODE allocations to fail
if the node has no memory.
Currently, alloc_fresh_huge_page() returns NULL when it is not able to
allocate a huge page on the current node, as specified by its custom
interleave variable. The callers of this function, though, assume that a
failure in alloc_fresh_huge_page() indicates no hugepages can be allocated
on the system period. This might not be the case, for instance, if we have
an uneven NUMA system, and we happen to try to allocate a hugepage on a
node with less memory and fail, while there is still plenty of free memory
on the other nodes.
To correct this, make alloc_fresh_huge_page() search through all online
nodes before deciding no hugepages can be allocated. Add a helper function
for actually allocating the hugepage. Use a new global nid iterator to
control which nid to allocate on.
Note: we expect particular semantics for __GFP_THISNODE, which are now
enforced even for memoryless nodes. That is, there is should be no
fallback to other nodes. Therefore, we rely on the nid passed into
alloc_pages_node() to be the nid the page comes from. If this is
incorrect, accounting will break.
Tested on x86 !NUMA, x86 NUMA, x86_64 NUMA and ppc64 NUMA (with 2
memoryless nodes).
Before on the ppc64 box:
Trying to clear the hugetlb pool
Done. 0 free
Trying to resize the pool to 100
Node 0 HugePages_Free: 25
Node 1 HugePages_Free: 75
Node 2 HugePages_Free: 0
Node 3 HugePages_Free: 0
Done. Initially 100 free
Trying to resize the pool to 200
Node 0 HugePages_Free: 50
Node 1 HugePages_Free: 150
Node 2 HugePages_Free: 0
Node 3 HugePages_Free: 0
Done. 200 free
After:
Trying to clear the hugetlb pool
Done. 0 free
Trying to resize the pool to 100
Node 0 HugePages_Free: 50
Node 1 HugePages_Free: 50
Node 2 HugePages_Free: 0
Node 3 HugePages_Free: 0
Done. Initially 100 free
Trying to resize the pool to 200
Node 0 HugePages_Free: 100
Node 1 HugePages_Free: 100
Node 2 HugePages_Free: 0
Node 3 HugePages_Free: 0
Done. 200 free
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Acked-by: Christoph Lameter <clameter@sgi.com>
Cc: Adam Litke <agl@us.ibm.com>
Cc: David Gibson <hermes@gibson.dropbear.id.au>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Ken Chen <kenchen@google.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
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>
When shrinking the size of the hugetlb pool via the nr_hugepages sysctl, we
are careful to keep enough pages around to satisfy reservations. But the
calculation is flawed for the following scenario:
Action Pool Counters (Total, Free, Resv)
====== =============
Set pool to 1 page 1 1 0
Map 1 page MAP_PRIVATE 1 1 0
Touch the page to fault it in 1 0 0
Set pool to 3 pages 3 2 0
Map 2 pages MAP_SHARED 3 2 2
Set pool to 2 pages 2 1 2 <-- Mistake, should be 3 2 2
Touch the 2 shared pages 2 0 1 <-- Program crashes here
The last touch above will terminate the process due to lack of huge pages.
This patch corrects the calculation so that it factors in pages being used
for private mappings. Andrew, this is a standalone fix suitable for
mainline. It is also now corrected in my latest dynamic pool resizing
patchset which I will send out soon.
Signed-off-by: Adam Litke <agl@us.ibm.com>
Acked-by: Ken Chen <kenchen@google.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The maximum size of the huge page pool can be controlled using the overall
size of the hugetlb filesystem (via its 'size' mount option). However in the
common case the this will not be set as the pool is traditionally fixed in
size at boot time. In order to maintain the expected semantics, we need to
prevent the pool expanding by default.
This patch introduces a new sysctl controlling dynamic pool resizing. When
this is enabled the pool will expand beyond its base size up to the size of
the hugetlb filesystem. It is disabled by default.
Signed-off-by: Adam Litke <agl@us.ibm.com>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Dave McCracken <dave.mccracken@oracle.com>
Cc: William Irwin <bill.irwin@oracle.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Ken Chen <kenchen@google.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Shared mappings require special handling because the huge pages needed to
fully populate the VMA must be reserved at mmap time. If not enough pages are
available when making the reservation, allocate all of the shortfall at once
from the buddy allocator and add the pages directly to the hugetlb pool. If
they cannot be allocated, then fail the mapping. The page surplus is
accounted for in the same way as for private mappings; faulted surplus pages
will be freed at unmap time. Reserved, surplus pages that have not been used
must be freed separately when their reservation has been released.
Signed-off-by: Adam Litke <agl@us.ibm.com>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Dave McCracken <dave.mccracken@oracle.com>
Cc: William Irwin <bill.irwin@oracle.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Ken Chen <kenchen@google.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Because we overcommit hugepages for MAP_PRIVATE mappings, it is possible that
the hugetlb pool will be exhausted or completely reserved when a hugepage is
needed to satisfy a page fault. Before killing the process in this situation,
try to allocate a hugepage directly from the buddy allocator.
The explicitly configured pool size becomes a low watermark. When dynamically
grown, the allocated huge pages are accounted as a surplus over the watermark.
As huge pages are freed on a node, surplus pages are released to the buddy
allocator so that the pool will shrink back to the watermark.
Surplus accounting also allows for friendlier explicit pool resizing. When
shrinking a pool that is fully in-use, increase the surplus so pages will be
returned to the buddy allocator as soon as they are freed. When growing a
pool that has a surplus, consume the surplus first and then allocate new
pages.
Signed-off-by: Adam Litke <agl@us.ibm.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Dave McCracken <dave.mccracken@oracle.com>
Cc: William Irwin <bill.irwin@oracle.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Ken Chen <kenchen@google.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dynamic huge page pool resizing.
In most real-world scenarios, configuring the size of the hugetlb pool
correctly is a difficult task. If too few pages are allocated to the pool,
applications using MAP_SHARED may fail to mmap() a hugepage region and
applications using MAP_PRIVATE may receive SIGBUS. Isolating too much memory
in the hugetlb pool means it is not available for other uses, especially those
programs not using huge pages.
The obvious answer is to let the hugetlb pool grow and shrink in response to
the runtime demand for huge pages. The work Mel Gorman has been doing to
establish a memory zone for movable memory allocations makes dynamically
resizing the hugetlb pool reliable within the limits of that zone. This patch
series implements dynamic pool resizing for private and shared mappings while
being careful to maintain existing semantics. Please reply with your comments
and feedback; even just to say whether it would be a useful feature to you.
Thanks.
How it works
============
Upon depletion of the hugetlb pool, rather than reporting an error immediately,
first try and allocate the needed huge pages directly from the buddy allocator.
Care must be taken to avoid unbounded growth of the hugetlb pool, so the
hugetlb filesystem quota is used to limit overall pool size.
The real work begins when we decide there is a shortage of huge pages. What
happens next depends on whether the pages are for a private or shared mapping.
Private mappings are straightforward. At fault time, if alloc_huge_page()
fails, we allocate a page from the buddy allocator and increment the source
node's surplus_huge_pages counter. When free_huge_page() is called for a page
on a node with a surplus, the page is freed directly to the buddy allocator
instead of the hugetlb pool.
Because shared mappings require all of the pages to be reserved up front, some
additional work must be done at mmap() to support them. We determine the
reservation shortage and allocate the required number of pages all at once.
These pages are then added to the hugetlb pool and marked reserved. Where that
is not possible the mmap() will fail. As with private mappings, the
appropriate surplus counters are updated. Since reserved huge pages won't
necessarily be used by the process, we can't be sure that free_huge_page() will
always be called to return surplus pages to the buddy allocator. To prevent
the huge page pool from bloating, we must free unused surplus pages when their
reservation has ended.
Controlling it
==============
With the entire patch series applied, pool resizing is off by default so unless
specific action is taken, the semantics are unchanged.
To take advantage of the flexibility afforded by this patch series one must
tolerate a change in semantics. To control hugetlb pool growth, the following
techniques can be employed:
* A sysctl tunable to enable/disable the feature entirely
* The size= mount option for hugetlbfs filesystems to limit pool size
Performance
===========
When contiguous memory is readily available, it is expected that the cost of
dynamicly resizing the pool will be small. This series has been performance
tested with 'stream' to measure this cost.
Stream (http://www.cs.virginia.edu/stream/) was linked with libhugetlbfs to
enable remapping of the text and data/bss segments into huge pages.
Stream with small array
-----------------------
Baseline: nr_hugepages = 0, No libhugetlbfs segment remapping
Preallocated: nr_hugepages = 5, Text and data/bss remapping
Dynamic: nr_hugepages = 0, Text and data/bss remapping
Rate (MB/s)
Function Baseline Preallocated Dynamic
Copy: 4695.6266 5942.8371 5982.2287
Scale: 4451.5776 5017.1419 5658.7843
Add: 5815.8849 7927.7827 8119.3552
Triad: 5949.4144 8527.6492 8110.6903
Stream with large array
-----------------------
Baseline: nr_hugepages = 0, No libhugetlbfs segment remapping
Preallocated: nr_hugepages = 67, Text and data/bss remapping
Dynamic: nr_hugepages = 0, Text and data/bss remapping
Rate (MB/s)
Function Baseline Preallocated Dynamic
Copy: 2227.8281 2544.2732 2546.4947
Scale: 2136.3208 2430.7294 2421.2074
Add: 2773.1449 4004.0021 3999.4331
Triad: 2748.4502 3777.0109 3773.4970
* All numbers are averages taken from 10 consecutive runs with a maximum
standard deviation of 1.3 percent noted.
This patch:
Simply move update_and_free_page() so that it can be reused later in this
patch series. The implementation is not changed.
Signed-off-by: Adam Litke <agl@us.ibm.com>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Dave McCracken <dave.mccracken@oracle.com>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Ken Chen <kenchen@google.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch is to avoid panic when memory hot-add is executed with
sparsemem-vmemmap. Current vmemmap-sparsemem code doesn't support memory
hot-add. Vmemmap must be populated when hot-add. This is for
2.6.23-rc2-mm2.
Todo: # Even if this patch is applied, the message "[xxxx-xxxx] potential
offnode page_structs" is displayed. To allocate memmap on its node,
memmap (and pgdat) must be initialized itself like chicken and
egg relationship.
# vmemmap_unpopulate will be necessary for followings.
- For cancel hot-add due to error.
- For unplug.
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Christoph Lameter <clameter@sgi.com>
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>
Now, arch dependent code around CONFIG_MEMORY_HOTREMOVE is a mess.
This patch cleans up them. This is against 2.6.23-rc6-mm1.
- fix compile failure on ia64/ CONFIG_MEMORY_HOTPLUG && !CONFIG_MEMORY_HOTREMOVE case.
- For !CONFIG_MEMORY_HOTREMOVE, add generic no-op remove_memory(),
which returns -EINVAL.
- removed remove_pages() only used in powerpc.
- removed no-op remove_memory() in i386, sh, sparc64, x86_64.
- only powerpc returns -ENOSYS at memory hot remove(no-op). changes it
to return -EINVAL.
Note:
Currently, only ia64 supports CONFIG_MEMORY_HOTREMOVE. I welcome other
archs if there are requirements and testers.
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>
Logic.
- set all pages in [start,end) as isolated migration-type.
by this, all free pages in the range will be not-for-use.
- Migrate all LRU pages in the range.
- Test all pages in the range's refcnt is zero or not.
Todo:
- allocate migration destination page from better area.
- confirm page_count(page)== 0 && PageReserved(page) page is safe to be freed..
(I don't like this kind of page but..
- Find out pages which cannot be migrated.
- more running tests.
- Use reclaim for unplugging other memory type area.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-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>
Implement generic chunk-of-pages isolation method by using page grouping ops.
This patch add MIGRATE_ISOLATE to MIGRATE_TYPES. By this
- MIGRATE_TYPES increases.
- bitmap for migratetype is enlarged.
pages of MIGRATE_ISOLATE migratetype will not be allocated even if it is free.
By this, you can isolated *freed* pages from users. How-to-free pages is not
a purpose of this patch. You may use reclaim and migrate codes to free pages.
If start_isolate_page_range(start,end) is called,
- migratetype of the range turns to be MIGRATE_ISOLATE if
its type is MIGRATE_MOVABLE. (*) this check can be updated if other
memory reclaiming works make progress.
- MIGRATE_ISOLATE is not on migratetype fallback list.
- All free pages and will-be-freed pages are isolated.
To check all pages in the range are isolated or not, use test_pages_isolated(),
To cancel isolation, use undo_isolate_page_range().
Changes V6 -> V7
- removed unnecessary #ifdef
There are HOLES_IN_ZONE handling codes...I'm glad if we can remove them..
Signed-off-by: Yasunori Goto <y-goto@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>
A clean up patch for "scanning memory resource [start, end)" operation.
Now, find_next_system_ram() function is used in memory hotplug, but this
interface is not easy to use and codes are complicated.
This patch adds walk_memory_resouce(start,len,arg,func) function.
The function 'func' is called per valid memory resouce range in [start,pfn).
[pbadari@us.ibm.com: Error handling in walk_memory_resource()]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The statistics patch later needs to know what order a free page is on the free
lists. Rather than having special knowledge of page_private() when
PageBuddy() is set, this patch places out page_order() in internal.h and adds
a VM_BUG_ON to catch using it on non-PageBuddy pages.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Christoph Lameter <clameter@sgi.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>
We touch a cacheline in the kmem_cache structure for zeroing to get the
size. However, the hot paths in slab_alloc and slab_free do not reference
any other fields in kmem_cache, so we may have to just bring in the
cacheline for this one access.
Add a new field to kmem_cache_cpu that contains the object size. That
cacheline must already be used in the hotpaths. So we save one cacheline
on every slab_alloc if we zero.
We need to update the kmem_cache_cpu object size if an aliasing operation
changes the objsize of an non debug slab.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The kmem_cache_cpu structures introduced are currently an array placed in the
kmem_cache struct. Meaning the kmem_cache_cpu structures are overwhelmingly
on the wrong node for systems with a higher amount of nodes. These are
performance critical structures since the per node information has
to be touched for every alloc and free in a slab.
In order to place the kmem_cache_cpu structure optimally we put an array
of pointers to kmem_cache_cpu structs in kmem_cache (similar to SLAB).
However, the kmem_cache_cpu structures can now be allocated in a more
intelligent way.
We would like to put per cpu structures for the same cpu but different
slab caches in cachelines together to save space and decrease the cache
footprint. However, the slab allocators itself control only allocations
per node. We set up a simple per cpu array for every processor with
100 per cpu structures which is usually enough to get them all set up right.
If we run out then we fall back to kmalloc_node. This also solves the
bootstrap problem since we do not have to use slab allocator functions
early in boot to get memory for the small per cpu structures.
Pro:
- NUMA aware placement improves memory performance
- All global structures in struct kmem_cache become readonly
- Dense packing of per cpu structures reduces cacheline
footprint in SMP and NUMA.
- Potential avoidance of exclusive cacheline fetches
on the free and alloc hotpath since multiple kmem_cache_cpu
structures are in one cacheline. This is particularly important
for the kmalloc array.
Cons:
- Additional reference to one read only cacheline (per cpu
array of pointers to kmem_cache_cpu) in both slab_alloc()
and slab_free().
[akinobu.mita@gmail.com: fix cpu hotplug offline/online path]
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: "Pekka Enberg" <penberg@cs.helsinki.fi>
Cc: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Set c->node to -1 if we allocate from a debug slab instead for SlabDebug
which requires access the page struct cacheline.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Tested-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We need the offset from the page struct during slab_alloc and slab_free. In
both cases we also reference the cacheline of the kmem_cache_cpu structure.
We can therefore move the offset field into the kmem_cache_cpu structure
freeing up 16 bits in the page struct.
Moving the offset allows an allocation from slab_alloc() without touching the
page struct in the hot path.
The only thing left in slab_free() that touches the page struct cacheline for
per cpu freeing is the checking of SlabDebug(page). The next patch deals with
that.
Use the available 16 bits to broaden page->inuse. More than 64k objects per
slab become possible and we can get rid of the checks for that limitation.
No need anymore to shrink the order of slabs if we boot with 2M sized slabs
(slub_min_order=9).
No need anymore to switch off the offset calculation for very large slabs
since the field in the kmem_cache_cpu structure is 32 bits and so the offset
field can now handle slab sizes of up to 8GB.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After moving the lockless_freelist to kmem_cache_cpu we no longer need
page->lockless_freelist. Restructure the use of the struct page fields in
such a way that we never touch the mapping field.
This is turn allows us to remove the special casing of SLUB when determining
the mapping of a page (needed for corner cases of virtual caches machines that
need to flush caches of processors mapping a page).
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
A remote free may access the same page struct that also contains the lockless
freelist for the cpu slab. If objects have a short lifetime and are freed by
a different processor then remote frees back to the slab from which we are
currently allocating are frequent. The cacheline with the page struct needs
to be repeately acquired in exclusive mode by both the allocating thread and
the freeing thread. If this is frequent enough then performance will suffer
because of cacheline bouncing.
This patchset puts the lockless_freelist pointer in its own cacheline. In
order to make that happen we introduce a per cpu structure called
kmem_cache_cpu.
Instead of keeping an array of pointers to page structs we now keep an array
to a per cpu structure that--among other things--contains the pointer to the
lockless freelist. The freeing thread can then keep possession of exclusive
access to the page struct cacheline while the allocating thread keeps its
exclusive access to the cacheline containing the per cpu structure.
This works as long as the allocating cpu is able to service its request
from the lockless freelist. If the lockless freelist runs empty then the
allocating thread needs to acquire exclusive access to the cacheline with
the page struct lock the slab.
The allocating thread will then check if new objects were freed to the per
cpu slab. If so it will keep the slab as the cpu slab and continue with the
recently remote freed objects. So the allocating thread can take a series
of just freed remote pages and dish them out again. Ideally allocations
could be just recycling objects in the same slab this way which will lead
to an ideal allocation / remote free pattern.
The number of objects that can be handled in this way is limited by the
capacity of one slab. Increasing slab size via slub_min_objects/
slub_max_order may increase the number of objects and therefore performance.
If the allocating thread runs out of objects and finds that no objects were
put back by the remote processor then it will retrieve a new slab (from the
partial lists or from the page allocator) and start with a whole
new set of objects while the remote thread may still be freeing objects to
the old cpu slab. This may then repeat until the new slab is also exhausted.
If remote freeing has freed objects in the earlier slab then that earlier
slab will now be on the partial freelist and the allocating thread will
pick that slab next for allocation. So the loop is extended. However,
both threads need to take the list_lock to make the swizzling via
the partial list happen.
It is likely that this kind of scheme will keep the objects being passed
around to a small set that can be kept in the cpu caches leading to increased
performance.
More code cleanups become possible:
- Instead of passing a cpu we can now pass a kmem_cache_cpu structure around.
Allows reducing the number of parameters to various functions.
- Can define a new node_match() function for NUMA to encapsulate locality
checks.
Effect on allocations:
Cachelines touched before this patch:
Write: page cache struct and first cacheline of object
Cachelines touched after this patch:
Write: kmem_cache_cpu cacheline and first cacheline of object
Read: page cache struct (but see later patch that avoids touching
that cacheline)
The handling when the lockless alloc list runs empty gets to be a bit more
complicated since another cacheline has now to be written to. But that is
halfway out of the hot path.
Effect on freeing:
Cachelines touched before this patch:
Write: page_struct and first cacheline of object
Cachelines touched after this patch depending on how we free:
Write(to cpu_slab): kmem_cache_cpu struct and first cacheline of object
Write(to other): page struct and first cacheline of object
Read(to cpu_slab): page struct to id slab etc. (but see later patch that
avoids touching the page struct on free)
Read(to other): cpu local kmem_cache_cpu struct to verify its not
the cpu slab.
Summary:
Pro:
- Distinct cachelines so that concurrent remote frees and local
allocs on a cpuslab can occur without cacheline bouncing.
- Avoids potential bouncing cachelines because of neighboring
per cpu pointer updates in kmem_cache's cpu_slab structure since
it now grows to a cacheline (Therefore remove the comment
that talks about that concern).
Cons:
- Freeing objects now requires the reading of one additional
cacheline. That can be mitigated for some cases by the following
patches but its not possible to completely eliminate these
references.
- Memory usage grows slightly.
The size of each per cpu object is blown up from one word
(pointing to the page_struct) to one cacheline with various data.
So this is NR_CPUS*NR_SLABS*L1_BYTES more memory use. Lets say
NR_SLABS is 100 and a cache line size of 128 then we have just
increased SLAB metadata requirements by 12.8k per cpu.
(Another later patch reduces these requirements)
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Fengguang Wu