Minchan Kim [Mon, 8 Oct 2012 23:32:54 +0000 (16:32 -0700)]
memory-hotplug: don't replace lowmem pages with highmem
The changelog for commit 6a6dccba2fdc ("mm: cma: don't replace lowmem
pages with highmem") mentioned that lowmem pages can be replaced by
highmem pages during CMA migration. 6a6dccba2fdc fixed that issue.
Quote from that changelog:
: The filesystem layer expects pages in the block device's mapping to not
: be in highmem (the mapping's gfp mask is set in bdget()), but CMA can
: currently replace lowmem pages with highmem pages, leading to crashes in
: filesystem code such as the one below:
:
: Unable to handle kernel NULL pointer dereference at virtual address 00000400
: pgd = c0c98000
: [00000400] *pgd=00c91831, *pte=00000000, *ppte=00000000
: Internal error: Oops: 817 [#1] PREEMPT SMP ARM
: CPU: 0 Not tainted (3.5.0-rc5+ #80)
: PC is at __memzero+0x24/0x80
: ...
: Process fsstress (pid: 323, stack limit = 0xc0cbc2f0)
: Backtrace:
: [<c010e3f0>] (ext4_getblk+0x0/0x180) from [<c010e58c>] (ext4_bread+0x1c/0x98)
: [<c010e570>] (ext4_bread+0x0/0x98) from [<c0117944>] (ext4_mkdir+0x160/0x3bc)
: r4:c15337f0
: [<c01177e4>] (ext4_mkdir+0x0/0x3bc) from [<c00c29e0>] (vfs_mkdir+0x8c/0x98)
: [<c00c2954>] (vfs_mkdir+0x0/0x98) from [<c00c2a60>] (sys_mkdirat+0x74/0xac)
: r6:00000000 r5:c152eb40 r4:000001ff r3:c14b43f0
: [<c00c29ec>] (sys_mkdirat+0x0/0xac) from [<c00c2ab8>] (sys_mkdir+0x20/0x24)
: r6:beccdcf0 r5:00074000 r4:beccdbbc
: [<c00c2a98>] (sys_mkdir+0x0/0x24) from [<c000e3c0>] (ret_fast_syscall+0x0/0x30)
Memory-hotplug has same problem as CMA has so the same fix can be applied
to memory-hotplug as well.
Fix it by reusing.
Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Acked-by: Michal Nazarewicz <mina86@mina86.com> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Wen Congyang <wency@cn.fujitsu.com> Acked-by: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:47 +0000 (16:32 -0700)]
mm: compaction: clear PG_migrate_skip based on compaction and reclaim activity
Compaction caches if a pageblock was scanned and no pages were isolated so
that the pageblocks can be skipped in the future to reduce scanning. This
information is not cleared by the page allocator based on activity due to
the impact it would have to the page allocator fast paths. Hence there is
a requirement that something clear the cache or pageblocks will be skipped
forever. Currently the cache is cleared if there were a number of recent
allocation failures and it has not been cleared within the last 5 seconds.
Time-based decisions like this are terrible as they have no relationship
to VM activity and is basically a big hammer.
Unfortunately, accurate heuristics would add cost to some hot paths so
this patch implements a rough heuristic. There are two cases where the
cache is cleared.
1. If a !kswapd process completes a compaction cycle (migrate and free
scanner meet), the zone is marked compact_blockskip_flush. When kswapd
goes to sleep, it will clear the cache. This is expected to be the
common case where the cache is cleared. It does not really matter if
kswapd happens to be asleep or going to sleep when the flag is set as
it will be woken on the next allocation request.
2. If there have been multiple failures recently and compaction just
finished being deferred then a process will clear the cache and start a
full scan. This situation happens if there are multiple high-order
allocation requests under heavy memory pressure.
The clearing of the PG_migrate_skip bits and other scans is inherently
racy but the race is harmless. For allocations that can fail such as THP,
they will simply fail. For requests that cannot fail, they will retry the
allocation. Tests indicated that scanning rates were roughly similar to
when the time-based heuristic was used and the allocation success rates
were similar.
Signed-off-by: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Richard Davies <richard@arachsys.com> Cc: Shaohua Li <shli@kernel.org> Cc: Avi Kivity <avi@redhat.com> Cc: Rafael Aquini <aquini@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:45 +0000 (16:32 -0700)]
mm: compaction: Restart compaction from near where it left off
This is almost entirely based on Rik's previous patches and discussions
with him about how this might be implemented.
Order > 0 compaction stops when enough free pages of the correct page
order have been coalesced. When doing subsequent higher order
allocations, it is possible for compaction to be invoked many times.
However, the compaction code always starts out looking for things to
compact at the start of the zone, and for free pages to compact things to
at the end of the zone.
This can cause quadratic behaviour, with isolate_freepages starting at the
end of the zone each time, even though previous invocations of the
compaction code already filled up all free memory on that end of the zone.
This can cause isolate_freepages to take enormous amounts of CPU with
certain workloads on larger memory systems.
This patch caches where the migration and free scanner should start from
on subsequent compaction invocations using the pageblock-skip information.
When compaction starts it begins from the cached restart points and will
update the cached restart points until a page is isolated or a pageblock
is skipped that would have been scanned by synchronous compaction.
Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Rik van Riel <riel@redhat.com> Cc: Richard Davies <richard@arachsys.com> Cc: Shaohua Li <shli@kernel.org> Cc: Avi Kivity <avi@redhat.com> Acked-by: Rafael Aquini <aquini@redhat.com> Cc: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:41 +0000 (16:32 -0700)]
mm: compaction: cache if a pageblock was scanned and no pages were isolated
When compaction was implemented it was known that scanning could
potentially be excessive. The ideal was that a counter be maintained for
each pageblock but maintaining this information would incur a severe
penalty due to a shared writable cache line. It has reached the point
where the scanning costs are a serious problem, particularly on
long-lived systems where a large process starts and allocates a large
number of THPs at the same time.
Instead of using a shared counter, this patch adds another bit to the
pageblock flags called PG_migrate_skip. If a pageblock is scanned by
either migrate or free scanner and 0 pages were isolated, the pageblock is
marked to be skipped in the future. When scanning, this bit is checked
before any scanning takes place and the block skipped if set.
The main difficulty with a patch like this is "when to ignore the cached
information?" If it's ignored too often, the scanning rates will still be
excessive. If the information is too stale then allocations will fail
that might have otherwise succeeded. In this patch
o CMA always ignores the information
o If the migrate and free scanner meet then the cached information will
be discarded if it's at least 5 seconds since the last time the cache
was discarded
o If there are a large number of allocation failures, discard the cache.
The time-based heuristic is very clumsy but there are few choices for a
better event. Depending solely on multiple allocation failures still
allows excessive scanning when THP allocations are failing in quick
succession due to memory pressure. Waiting until memory pressure is
relieved would cause compaction to continually fail instead of using
reclaim/compaction to try allocate the page. The time-based mechanism is
clumsy but a better option is not obvious.
Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Rik van Riel <riel@redhat.com> Cc: Richard Davies <richard@arachsys.com> Cc: Shaohua Li <shli@kernel.org> Cc: Avi Kivity <avi@redhat.com> Acked-by: Rafael Aquini <aquini@redhat.com> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Cc: Kyungmin Park <kyungmin.park@samsung.com> Cc: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:40 +0000 (16:32 -0700)]
revert "mm: have order > 0 compaction start off where it left"
This reverts commit 7db8889ab05b ("mm: have order > 0 compaction start
off where it left") and commit de74f1cc ("mm: have order > 0 compaction
start near a pageblock with free pages"). These patches were a good
idea and tests confirmed that they massively reduced the amount of
scanning but the implementation is complex and tricky to understand. A
later patch will cache what pageblocks should be skipped and
reimplements the concept of compact_cached_free_pfn on top for both
migration and free scanners.
Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Rik van Riel <riel@redhat.com> Cc: Richard Davies <richard@arachsys.com> Cc: Shaohua Li <shli@kernel.org> Cc: Avi Kivity <avi@redhat.com> Acked-by: Rafael Aquini <aquini@redhat.com> Acked-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:36 +0000 (16:32 -0700)]
mm: compaction: acquire the zone->lock as late as possible
Compaction's free scanner acquires the zone->lock when checking for
PageBuddy pages and isolating them. It does this even if there are no
PageBuddy pages in the range.
This patch defers acquiring the zone lock for as long as possible. In the
event there are no free pages in the pageblock then the lock will not be
acquired at all which reduces contention on zone->lock.
Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Rik van Riel <riel@redhat.com> Cc: Richard Davies <richard@arachsys.com> Cc: Shaohua Li <shli@kernel.org> Cc: Avi Kivity <avi@redhat.com> Acked-by: Rafael Aquini <aquini@redhat.com> Acked-by: Minchan Kim <minchan@kernel.org> Tested-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:33 +0000 (16:32 -0700)]
mm: compaction: acquire the zone->lru_lock as late as possible
Richard Davies and Shaohua Li have both reported lock contention problems
in compaction on the zone and LRU locks as well as significant amounts of
time being spent in compaction. This series aims to reduce lock
contention and scanning rates to reduce that CPU usage. Richard reported
at https://lkml.org/lkml/2012/9/21/91 that this series made a big
different to a problem he reported in August:
Patch 1 defers acquiring the zone->lru_lock as long as possible.
Patch 2 defers acquiring the zone->lock as lock as possible.
Patch 3 reverts Rik's "skip-free" patches as the core concept gets
reimplemented later and the remaining patches are easier to
understand if this is reverted first.
Patch 4 adds a pageblock-skip bit to the pageblock flags to cache what
pageblocks should be skipped by the migrate and free scanners.
This drastically reduces the amount of scanning compaction has
to do.
Patch 5 reimplements something similar to Rik's idea except it uses the
pageblock-skip information to decide where the scanners should
restart from and does not need to wrap around.
I tested this on 3.6-rc6 + linux-next/akpm. Kernels tested were
Stress high-order allocation tests looked ok. Success rates are more or
less the same with the full series applied but there is an expectation
that there is less opportunity to race with other allocation requests if
there is less scanning. The time to complete the tests did not vary that
much and are uninteresting as were the vmstat statistics so I will not
present them here.
Using ftrace I recorded how much scanning was done by compaction and got this
The efficiency is worthless because of the nature of the test and the
number of failures. The really interesting point as far as this patch
series is concerned is the number of pages scanned. Note that reverting
Rik's patches massively increases the number of pages scanned indicating
that those patches really did make a difference to CPU usage.
However, caching what pageblocks should be skipped has a much higher
impact. With patches 1-8 applied, free page and migrate page scanning are
both reduced by 95% in comparison to the akpm kernel. If the basic
concept of Rik's patches are implemened on top then scanning then the free
scanner barely changed but migrate scanning was further reduced. That
said, tests on 3.6-rc5 indicated that the last patch had greater impact
than what was measured here so it is a bit variable.
One way or the other, this series has a large impact on the amount of
scanning compaction does when there is a storm of THP allocations.
This patch:
Compaction's migrate scanner acquires the zone->lru_lock when scanning a
range of pages looking for LRU pages to acquire. It does this even if
there are no LRU pages in the range. If multiple processes are compacting
then this can cause severe locking contention. To make matters worse
commit b2eef8c0 ("mm: compaction: minimise the time IRQs are disabled
while isolating pages for migration") releases the lru_lock every
SWAP_CLUSTER_MAX pages that are scanned.
This patch makes two changes to how the migrate scanner acquires the LRU
lock. First, it only releases the LRU lock every SWAP_CLUSTER_MAX pages
if the lock is contended. This reduces the number of times it
unnecessarily disables and re-enables IRQs. The second is that it defers
acquiring the LRU lock for as long as possible. If there are no LRU pages
or the only LRU pages are transhuge then the LRU lock will not be acquired
at all which reduces contention on zone->lru_lock.
[minchan@kernel.org: augment comment]
[akpm@linux-foundation.org: tweak comment text] Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Rik van Riel <riel@redhat.com> Cc: Richard Davies <richard@arachsys.com> Cc: Shaohua Li <shli@kernel.org> Cc: Avi Kivity <avi@redhat.com> Acked-by: Rafael Aquini <aquini@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:32:30 +0000 (16:32 -0700)]
mm: compaction: move fatal signal check out of compact_checklock_irqsave
Commit c67fe3752abe ("mm: compaction: Abort async compaction if locks
are contended or taking too long") addressed a lock contention problem
in compaction by introducing compact_checklock_irqsave() that effecively
aborting async compaction in the event of compaction.
To preserve existing behaviour it also moved a fatal_signal_pending()
check into compact_checklock_irqsave() but that is very misleading. It
"hides" the check within a locking function but has nothing to do with
locking as such. It just happens to work in a desirable fashion.
This patch moves the fatal_signal_pending() check to
isolate_migratepages_range() where it belongs. Arguably the same check
should also happen when isolating pages for freeing but it's overkill.
Signed-off-by: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Shaohua Li [Mon, 8 Oct 2012 23:32:27 +0000 (16:32 -0700)]
mm: compaction: abort compaction loop if lock is contended or run too long
isolate_migratepages_range() might isolate no pages if for example when
zone->lru_lock is contended and running asynchronous compaction. In this
case, we should abort compaction, otherwise, compact_zone will run a
useless loop and make zone->lru_lock is even contended.
An additional check is added to ensure that cc.migratepages and
cc.freepages get properly drained whan compaction is aborted.
[minchan@kernel.org: Putback pages isolated for migration if aborting]
[akpm@linux-foundation.org: compact_zone_order requires non-NULL arg contended]
[akpm@linux-foundation.org: make compact_zone_order() require non-NULL arg `contended']
[minchan@kernel.org: Putback pages isolated for migration if aborting] Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Shaohua Li <shli@fusionio.com> Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Shaohua Li [Mon, 8 Oct 2012 23:32:19 +0000 (16:32 -0700)]
readahead: fault retry breaks mmap file read random detection
.fault now can retry. The retry can break state machine of .fault. In
filemap_fault, if page is miss, ra->mmap_miss is increased. In the second
try, since the page is in page cache now, ra->mmap_miss is decreased. And
these are done in one fault, so we can't detect random mmap file access.
Add a new flag to indicate .fault is tried once. In the second try, skip
ra->mmap_miss decreasing. The filemap_fault state machine is ok with it.
I only tested x86, didn't test other archs, but looks the change for other
archs is obvious, but who knows :)
Signed-off-by: Shaohua Li <shaohua.li@fusionio.com> Cc: Rik van Riel <riel@redhat.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The x86 implementation of atomic_dec_if_positive is quite generic, so make
it available to all architectures.
This is needed for "swap: add a simple detector for inappropriate swapin
readahead".
[akpm@linux-foundation.org: do the "#define foo foo" trick in the conventional manner] Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: "David S. Miller" <davem@davemloft.net> Cc: Rik van Riel <riel@redhat.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Michal Simek <monstr@monstr.eu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Minchan Kim [Mon, 8 Oct 2012 23:32:16 +0000 (16:32 -0700)]
memory-hotplug: fix pages missed by race rather than failing
If race between allocation and isolation in memory-hotplug offline
happens, some pages could be in MIGRATE_MOVABLE of free_list although the
pageblock's migratetype of the page is MIGRATE_ISOLATE.
The race could be detected by get_freepage_migratetype in
__test_page_isolated_in_pageblock. If it is detected, now EBUSY gets
bubbled all the way up and the hotplug operations fails.
But better idea is instead of returning and failing memory-hotremove, move
the free page to the correct list at the time it is detected. It could
enhance memory-hotremove operation success ratio although the race is
really rare.
free_hot_cold_page(Page A)
/* without zone->lock */
migratetype = get_pageblock_migratetype(Page A);
/*
* Page could be moved into MIGRATE_MOVABLE
* of per_cpu_pages
*/
list_add_tail(&page->lru, &pcp->lists[migratetype]);
/* Page A could be in MIGRATE_MOVABLE of free_list. */
check_pages_isolated
__test_page_isolated_in_pageblock
/*
* We can't catch freed page which
* is free_list[MIGRATE_MOVABLE]
*/
if (PageBuddy(page A))
pfn += 1 << page_order(page A);
/* So, Page A could be allocated */
__offline_isolated_pages
/*
* BUG_ON hit or offline page
* which is used by someone
*/
BUG_ON(!PageBuddy(page A));
This patch checks page's migratetype in freelist in
__test_page_isolated_in_pageblock. So now
__test_page_isolated_in_pageblock can check the page caused by above race
and can fail of memory offlining.
Minchan Kim [Mon, 8 Oct 2012 23:32:11 +0000 (16:32 -0700)]
mm: remain migratetype in freed page
The page allocator caches the pageblock information in page->private while
it is in the PCP freelists but this is overwritten with the order of the
page when freed to the buddy allocator. This patch stores the migratetype
of the page in the page->index field so that it is available at all times
when the page remain in free_list.
This patch adds a new call site in __free_pages_ok so it might be overhead
a bit but it's for high order allocation. So I believe damage isn't hurt.
Minchan Kim [Mon, 8 Oct 2012 23:32:08 +0000 (16:32 -0700)]
mm: page_alloc: use get_freepage_migratetype() instead of page_private()
The page allocator uses set_page_private and page_private for handling
migratetype when it frees page. Let's replace them with [set|get]
_freepage_migratetype to make it more clear.
* Add ALLOC_CMA alloc flag and pass it to [__]zone_watermark_ok()
(from Minchan Kim).
* During watermark check decrease available free pages number by
free CMA pages number if necessary (unmovable allocations cannot
use pages from CMA areas).
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add NR_FREE_CMA_PAGES counter to be later used for checking watermark in
__zone_watermark_ok(). For simplicity and to avoid #ifdef hell make this
counter always available (not only when CONFIG_CMA=y).
[akpm@linux-foundation.org: use conventional migratetype naming] Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Isolated free pages shouldn't be accounted to NR_FREE_PAGES counter. Fix
it by properly decreasing/increasing NR_FREE_PAGES counter in
set_migratetype_isolate()/unset_migratetype_isolate() and removing counter
adjustment for isolated pages from free_one_page() and split_free_page().
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
page->private gets re-used in __free_one_page() to store page order
(so trace_mm_page_pcpu_drain() may print order instead of migratetype)
thus migratetype value must be cached locally.
Fixes regression introduced in commit a7016235a61d ("mm: fix migratetype
bug which slowed swapping"). This caused incorrect data to be attached
to the mm_page_pcpu_drain trace event.
[akpm@linux-foundation.org: add comment] Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Michal Nazarewicz <mina86@mina86.com> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Mel Gorman <mgorman@suse.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Acked-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>
Minchan Kim [Mon, 8 Oct 2012 23:31:55 +0000 (16:31 -0700)]
mm: cma: discard clean pages during contiguous allocation instead of migration
Drop clean cache pages instead of migration during alloc_contig_range() to
minimise allocation latency by reducing the amount of migration that is
necessary. It's useful for CMA because latency of migration is more
important than evicting the background process's working set. In
addition, as pages are reclaimed then fewer free pages for migration
targets are required so it avoids memory reclaiming to get free pages,
which is a contributory factor to increased latency.
I measured elapsed time of __alloc_contig_migrate_range() which migrates
10M in 40M movable zone in QEMU machine.
Before - 146ms, After - 7ms
[akpm@linux-foundation.org: fix nommu build] Signed-off-by: Mel Gorman <mgorman@suse.de> Signed-off-by: Minchan Kim <minchan@kernel.org> Reviewed-by: Mel Gorman <mgorman@suse.de> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Acked-by: Michal Nazarewicz <mina86@mina86.com> Cc: Rik van Riel <riel@redhat.com> Tested-by: Kyungmin Park <kyungmin.park@samsung.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Xishi Qiu [Mon, 8 Oct 2012 23:31:51 +0000 (16:31 -0700)]
memory-hotplug: build zonelists when offlining pages
online_pages() does build_all_zonelists() and zone_pcp_update(), I think
offline_pages() should do it too.
When the zone has no memory to allocate, remove it from other nodes'
zonelists. zone_batchsize() depends on zone's present pages, if zone's
present pages are changed, zone's pcp should be updated.
During mremap(), the destination VMA is generally placed after the
original vma in rmap traversal order: in move_vma(), we always have
new_pgoff >= vma->vm_pgoff, and as a result new_vma->vm_pgoff >=
vma->vm_pgoff unless vma_merge() merged the new vma with an adjacent one.
When the destination VMA is placed after the original in rmap traversal
order, we can avoid taking the rmap locks in move_ptes().
Essentially, this reintroduces the optimization that had been disabled in
"mm anon rmap: remove anon_vma_moveto_tail". The difference is that we
don't try to impose the rmap traversal order; instead we just rely on
things being in the desired order in the common case and fall back to
taking locks in the uncommon case. Also we skip the i_mmap_mutex in
addition to the anon_vma lock: in both cases, the vmas are traversed in
increasing vm_pgoff order with ties resolved in tree insertion order.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add a CONFIG_DEBUG_VM_RB build option for the previously existing
DEBUG_MM_RB code. Now that Andi Kleen modified it to avoid using
recursive algorithms, we can expose it a bit more.
Also extend this code to validate_mm() after stack expansion, and to check
that the vma's start and last pgoffs have not changed since the nodes were
inserted on the anon vma interval tree (as it is important that the nodes
be reindexed after each such update).
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm rmap: remove vma_address check for address inside vma
In file and anon rmap, we use interval trees to find potentially relevant
vmas and then call vma_address() to find the virtual address the given
page might be found at in these vmas. vma_address() used to include a
check that the returned address falls within the limits of the vma, but
this check isn't necessary now that we always use interval trees in rmap:
the interval tree just doesn't return any vmas which this check would find
to be irrelevant. As a result, we can replace the use of -EFAULT error
code (which then needed to be checked in every call site) with a
VM_BUG_ON().
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm anon rmap: replace same_anon_vma linked list with an interval tree.
When a large VMA (anon or private file mapping) is first touched, which
will populate its anon_vma field, and then split into many regions through
the use of mprotect(), the original anon_vma ends up linking all of the
vmas on a linked list. This can cause rmap to become inefficient, as we
have to walk potentially thousands of irrelevent vmas before finding the
one a given anon page might fall into.
By replacing the same_anon_vma linked list with an interval tree (where
each avc's interval is determined by its vma's start and last pgoffs), we
can make rmap efficient for this use case again.
While the change is large, all of its pieces are fairly simple.
Most places that were walking the same_anon_vma list were looking for a
known pgoff, so they can just use the anon_vma_interval_tree_foreach()
interval tree iterator instead. The exception here is ksm, where the
page's index is not known. It would probably be possible to rework ksm so
that the index would be known, but for now I have decided to keep things
simple and just walk the entirety of the interval tree there.
When updating vma's that already have an anon_vma assigned, we must take
care to re-index the corresponding avc's on their interval tree. This is
done through the use of anon_vma_interval_tree_pre_update_vma() and
anon_vma_interval_tree_post_update_vma(), which remove the avc's from
their interval tree before the update and re-insert them after the update.
The anon_vma stays locked during the update, so there is no chance that
rmap would miss the vmas that are being updated.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mremap() had a clever optimization where move_ptes() did not take the
anon_vma lock to avoid a race with anon rmap users such as page migration.
Instead, the avc's were ordered in such a way that the origin vma was
always visited by rmap before the destination. This ordering and the use
of page table locks rmap usage safe. However, we want to replace the use
of linked lists in anon rmap with an interval tree, and this will make it
harder to impose such ordering as the interval tree will always be sorted
by the avc->vma->vm_pgoff value. For now, let's replace the
anon_vma_moveto_tail() ordering function with proper anon_vma locking in
move_ptes(). Once we have the anon interval tree in place, we will
re-introduce an optimization to avoid taking these locks in the most
common cases.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Update the generic interval tree code that was introduced in "mm: replace
vma prio_tree with an interval tree".
Changes:
- fixed 'endpoing' typo noticed by Andrew Morton
- replaced include/linux/interval_tree_tmpl.h, which was used as a
template (including it automatically defined the interval tree
functions) with include/linux/interval_tree_generic.h, which only
defines a preprocessor macro INTERVAL_TREE_DEFINE(), which itself
defines the interval tree functions when invoked. Now that is a very
long macro which is unfortunate, but it does make the usage sites
(lib/interval_tree.c and mm/interval_tree.c) a bit nicer than previously.
- make use of RB_DECLARE_CALLBACKS() in the INTERVAL_TREE_DEFINE() macro,
instead of duplicating that code in the interval tree template.
- replaced vma_interval_tree_add(), which was actually handling the
nonlinear and interval tree cases, with vma_interval_tree_insert_after()
which handles only the interval tree case and has an API that is more
consistent with the other interval tree handling functions.
The nonlinear case is now handled explicitly in kernel/fork.c dup_mmap().
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: move augmented rbtree functionality to rbtree_augmented.h
Provide rb_insert_augmented() and rb_erase_augmented() through a new
rbtree_augmented.h include file. rb_erase_augmented() is defined there as
an __always_inline function, in order to allow inlining of augmented
rbtree callbacks into it. Since this generates a relatively large
function, each augmented rbtree user should make sure to have a single
call site.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Hillf Danton <dhillf@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kmemleak uses a tree where each node represents an allocated memory object
in order to quickly find out what object a given address is part of.
However, the objects don't overlap, so rbtrees are a better choice than
prio tree for this use. They are both faster and have lower memory
overhead.
Tested by booting a kernel with kmemleak enabled, loading the
kmemleak_test module, and looking for the expected messages.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Hillf Danton <dhillf@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Tested-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Implement an interval tree as a replacement for the VMA prio_tree. The
algorithms are similar to lib/interval_tree.c; however that code can't be
directly reused as the interval endpoints are not explicitly stored in the
VMA. So instead, the common algorithm is moved into a template and the
details (node type, how to get interval endpoints from the node, etc) are
filled in using the C preprocessor.
Once the interval tree functions are available, using them as a
replacement to the VMA prio tree is a relatively simple, mechanical job.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Hillf Danton <dhillf@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch 1 implements support for interval trees, on top of the augmented
rbtree API. It also adds synthetic tests to compare the performance of
interval trees vs prio trees. Short answers is that interval trees are
slightly faster (~25%) on insert/erase, and much faster (~2.4 - 3x)
on search. It is debatable how realistic the synthetic test is, and I have
not made such measurements yet, but my impression is that interval trees
would still come out faster.
Patch 2 uses a preprocessor template to make the interval tree generic,
and uses it as a replacement for the vma prio_tree.
Patch 3 takes the other prio_tree user, kmemleak, and converts it to use
a basic rbtree. We don't actually need the augmented rbtree support here
because the intervals are always non-overlapping.
Patch 4 removes the now-unused prio tree library.
Patch 5 proposes an additional optimization to rb_erase_augmented, now
providing it as an inline function so that the augmented callbacks can be
inlined in. This provides an additional 5-10% performance improvement
for the interval tree insert/erase benchmark. There is a maintainance cost
as it exposes augmented rbtree users to some of the rbtree library internals;
however I think this cost shouldn't be too high as I expect the augmented
rbtree will always have much less users than the base rbtree.
I should probably add a quick summary of why I think it makes sense to
replace prio trees with augmented rbtree based interval trees now. One of
the drivers is that we need augmented rbtrees for Rik's vma gap finding
code, and once you have them, it just makes sense to use them for interval
trees as well, as this is the simpler and more well known algorithm. prio
trees, in comparison, seem *too* clever: they impose an additional 'heap'
constraint on the tree, which they use to guarantee a faster worst-case
complexity of O(k+log N) for stabbing queries in a well-balanced prio
tree, vs O(k*log N) for interval trees (where k=number of matches,
N=number of intervals). Now this sounds great, but in practice prio trees
don't realize this theorical benefit. First, the additional constraint
makes them harder to update, so that the kernel implementation has to
simplify things by balancing them like a radix tree, which is not always
ideal. Second, the fact that there are both index and heap properties
makes both tree manipulation and search more complex, which results in a
higher multiplicative time constant. As it turns out, the simple interval
tree algorithm ends up running faster than the more clever prio tree.
This patch:
Add two test modules:
- prio_tree_test measures the performance of lib/prio_tree.c, both for
insertion/removal and for stabbing searches
- interval_tree_test measures the performance of a library of equivalent
functionality, built using the augmented rbtree support.
In order to support the second test module, lib/interval_tree.c is
introduced. It is kept separate from the interval_tree_test main file
for two reasons: first we don't want to provide an unfair advantage
over prio_tree_test by having everything in a single compilation unit,
and second there is the possibility that the interval tree functionality
could get some non-test users in kernel over time.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Hillf Danton <dhillf@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
As proposed by Peter Zijlstra, this makes it easier to define the augmented
rbtree callbacks.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
convert arch/x86/mm/pat_rbtree.c to the proposed augmented rbtree api
and remove the old augmented rbtree implementation.
Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Introduce new augmented rbtree APIs that allow minimal recalculation of
augmented node information.
A new callback is added to the rbtree insertion and erase rebalancing
functions, to be called on each tree rotations. Such rotations preserve
the subtree's root augmented value, but require recalculation of the one
child that was previously located at the subtree root.
In the insertion case, the handcoded search phase must be updated to
maintain the augmented information on insertion, and then the rbtree
coloring/rebalancing algorithms keep it up to date.
In the erase case, things are more complicated since it is library
code that manipulates the rbtree in order to remove internal nodes.
This requires a couple additional callbacks to copy a subtree's
augmented value when a new root is stitched in, and to recompute
augmented values down the ancestry path when a node is removed from
the tree.
In order to preserve maximum speed for the non-augmented case,
we provide two versions of each tree manipulation function.
rb_insert_augmented() is the augmented equivalent of rb_insert_color(),
and rb_erase_augmented() is the augmented equivalent of rb_erase().
Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Small test to measure the performance of augmented rbtrees.
Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Various minor optimizations in rb_erase():
- Avoid multiple loading of node->__rb_parent_color when computing parent
and color information (possibly not in close sequence, as there might
be further branches in the algorithm)
- In the 1-child subcase of case 1, copy the __rb_parent_color field from
the erased node to the child instead of recomputing it from the desired
parent and color
- When searching for the erased node's successor, differentiate between
cases 2 and 3 based on whether any left links were followed. This avoids
a condition later down.
- In case 3, keep a pointer to the erased node's right child so we don't
have to refetch it later to adjust its parent.
- In the no-childs subcase of cases 2 and 3, place the rebalance assigment
last so that the compiler can remove the following if(rebalance) test.
Also, added some comments to illustrate cases 2 and 3.
Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: handle 1-child recoloring in rb_erase() instead of rb_erase_color()
An interesting observation for rb_erase() is that when a node has
exactly one child, the node must be black and the child must be red.
An interesting consequence is that removing such a node can be done by
simply replacing it with its child and making the child black,
which we can do efficiently in rb_erase(). __rb_erase_color() then
only needs to handle the no-childs case and can be modified accordingly.
Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In rb_erase, move the easy case (node to erase has no more than
1 child) first. I feel the code reads easier that way.
Signed-off-by: Michel Lespinasse <walken@google.com> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add __rb_change_child() as an inline helper function to replace code that
would otherwise be duplicated 4 times in the source.
No changes to binary size or speed.
Signed-off-by: Michel Lespinasse <walken@google.com> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When looking to fetch a node's sibling, we went through a sequence of:
- check if node is the parent's left child
- if it is, then fetch the parent's right child
This can be replaced with:
- fetch the parent's right child as an assumed sibling
- check that node is NOT the fetched child
This avoids fetching the parent's left child when node is actually
that child. Saves a bit on code size, though it doesn't seem to make
a large difference in speed.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: low level optimizations in __rb_erase_color()
In __rb_erase_color(), we often already have pointers to the nodes being
rotated and/or know what their colors must be, so we can generate more
efficient code than the generic __rb_rotate_left() and __rb_rotate_right()
functions.
Also when the current node is red or when flipping the sibling's color,
the parent is already known so we can use the more efficient
rb_set_parent_color() function to set the desired color.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: optimize case selection logic in __rb_erase_color()
In __rb_erase_color(), we have to select one of 3 cases depending on the
color on the 'other' node children. If both children are black, we flip a
few node colors and iterate. Otherwise, we do either one or two tree
rotations, depending on the color of the 'other' child opposite to 'node',
and then we are done.
The corresponding logic had duplicate checks for the color of the 'other'
child opposite to 'node'. It was checking it first to determine if both
children are black, and then to determine how many tree rotations are
required. Rearrange the logic to avoid that extra check.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: adjust node color in __rb_erase_color() only when necessary
In __rb_erase_color(), we were always setting a node to black after
exiting the main loop. And in one case, after fixing up the tree to
satisfy all rbtree invariants, we were setting the current node to root
just to guarantee a loop exit, at which point the root would be set to
black. However this is not necessary, as the root of an rbtree is already
known to be black. The only case where the color flip is required is when
we exit the loop due to the current node being red, and it's easiest to
just do the flip at that point instead of doing it after the loop.
[adrian.hunter@intel.com: perf tools: fix build for another rbtree.c change] Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: low level optimizations in rb_insert_color()
- Use the newly introduced rb_set_parent_color() function to flip the color
of nodes whose parent is already known.
- Optimize rb_parent() when the node is known to be red - there is no need
to mask out the color in that case.
- Flipping gparent's color to red requires us to fetch its rb_parent_color
field, so we can reuse it as the parent value for the next loop iteration.
- Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree
rotations: we already have pointers to all relevant nodes, and know their
colors (either because we want to adjust it, or because we've tested it,
or we can deduce it as black due to the node proximity to a known red node).
So we can generate more efficient code by making use of the node pointers
we already have, and setting both the parent and color attributes for
nodes all at once. Also in Case 2, some node attributes don't have to
be set because we know another tree rotation (Case 3) will always follow
and override them.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: adjust root color in rb_insert_color() only when necessary
The root node of an rbtree must always be black. However,
rb_insert_color() only needs to maintain this invariant when it has been
broken - that is, when it exits the loop due to the current (red) node
being the root. In all other cases (exiting after tree rotations, or
exiting due to an existing black parent) the invariant is already
satisfied, so there is no need to adjust the root node color.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: break out of rb_insert_color loop after tree rotation
It is a well known property of rbtrees that insertion never requires more
than two tree rotations. In our implementation, after one loop iteration
identified one or two necessary tree rotations, we would iterate and look
for more. However at that point the node's parent would always be black,
which would cause us to exit the loop.
We can make the code flow more obvious by just adding a break statement
after the tree rotations, where we know we are done. Additionally, in the
cases where two tree rotations are necessary, we don't have to update the
'node' pointer as it wouldn't be used until the next loop iteration, which
we now avoid due to this break statement.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This small module helps measure the performance of rbtree insert and
erase.
Additionally, we run a few correctness tests to check that the rbtrees
have all desired properties:
- contains the right number of nodes in the order desired,
- never two consecutive red nodes on any path,
- all paths to leaf nodes have the same number of black nodes,
- root node is black
[akpm@linux-foundation.org: fix printk warning: sparc64 cycles_t is unsigned long] Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: move some implementation details from rbtree.h to rbtree.c
rbtree users must use the documented APIs to manipulate the tree
structure. Low-level helpers to manipulate node colors and parenthood are
not part of that API, so move them to lib/rbtree.c
[dwmw2@infradead.org: fix jffs2 build issue due to renamed __rb_parent_color field] Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: fix incorrect rbtree node insertion in fs/proc/proc_sysctl.c
The recently added code to use rbtrees in sysctl did not follow the proper
rbtree interface on insertion - it was calling rb_link_node() which
inserts a new node into the binary tree, but missed the call to
rb_insert_color() which properly balances the rbtree and establishes all
expected rbtree invariants.
I found out about this only because faulty commit also used
rb_init_node(), which I am removing within this patchset. But I think
it's an easy mistake to make, and it makes me wonder if we should change
the rbtree API so that insertions would be done with a single rb_insert()
call (even if its implementation could still inline the rb_link_node()
part and call a private __rb_insert_color function to do the rebalancing).
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Empty nodes have no color. We can make use of this property to simplify
the code emitted by the RB_EMPTY_NODE and RB_CLEAR_NODE macros. Also,
we can get rid of the rb_init_node function which had been introduced by
commit 88d19cf37952 ("timers: Add rb_init_node() to allow for stack
allocated rb nodes") to avoid some issue with the empty node's color not
being initialized.
I'm not sure what the RB_EMPTY_NODE checks in rb_prev() / rb_next() are
doing there, though. axboe introduced them in commit 10fd48f2376d
("rbtree: fixed reversed RB_EMPTY_NODE and rb_next/prev"). The way I
see it, the 'empty node' abstraction is only used by rbtree users to
flag nodes that they haven't inserted in any rbtree, so asking the
predecessor or successor of such nodes doesn't make any sense.
One final rb_init_node() caller was recently added in sysctl code to
implement faster sysctl name lookups. This code doesn't make use of
RB_EMPTY_NODE at all, and from what I could see it only called
rb_init_node() under the mistaken assumption that such initialization was
required before node insertion.
[sfr@canb.auug.org.au: fix net/ceph/osd_client.c build] Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rbtree: reference Documentation/rbtree.txt for usage instructions
I recently started looking at the rbtree code (with an eye towards
improving the augmented rbtree support, but I haven't gotten there yet).
I noticed a lot of possible speed improvements, which I am now proposing
in this patch set.
Patches 1-4 are preparatory: remove internal functions from rbtree.h so
that users won't be tempted to use them instead of the documented APIs,
clean up some incorrect usages I've noticed (in particular, with the
recently added fs/proc/proc_sysctl.c rbtree usage), reference the
documentation so that people have one less excuse to miss it, etc.
Patch 5 is a small module I wrote to check the rbtree performance. It
creates 100 nodes with random keys and repeatedly inserts and erases them
from an rbtree. Additionally, it has code to check for rbtree invariants
after each insert or erase operation.
Patches 6-12 is where the rbtree optimizations are done, and they touch
only that one file, lib/rbtree.c . I am getting good results out of these
- in my small benchmark doing rbtree insertion (including search) and
erase, I'm seeing a 30% runtime reduction on Sandybridge E5, which is more
than I initially thought would be possible. (the results aren't as
impressive on my two other test hosts though, AMD barcelona and Intel
Westmere, where I am seeing 14% runtime reduction only). The code size -
both source (ommiting comments) and compiled - is also shorter after these
changes. However, I do admit that the updated code is more arduous to
read - one big reason for that is the removal of the tree rotation
helpers, which added some overhead but also made it easier to reason about
things locally. Overall, I believe this is an acceptable compromise,
given that this code doesn't get modified very often, and that I have good
tests for it.
Upon Peter's suggestion, I added comments showing the rtree configuration
before every rotation. I think they help; however it's still best to have
a copy of the cormen/leiserson/rivest book when digging into this code.
This patch: reference Documentation/rbtree.txt for usage instructions
include/linux/rbtree.h included some basic usage instructions, while
Documentation/rbtree.txt had some more complete and easier to follow
instructions. Replacing the former with a reference to the latter.
Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit d6629859b36d ("ipc/mqueue: improve performance of send/recv") and ce2d52cc ("ipc/mqueue: add rbtree node caching support") introduced an
rbtree of message priorities, and usage of rb_init_node() to initialize
the corresponding nodes. As it turns out, rb_init_node() is unnecessary
here, as the nodes are fully initialized on insertion by rb_link_node()
and the code doesn't access nodes that aren't inserted on the rbtree.
Removing the rb_init_node() calls as I removed that function during
rbtree API cleanups (the only other use of it was in a place that
similarly didn't require it).
Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Doug Ledford <dledford@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Gerald Schaefer [Mon, 8 Oct 2012 23:30:21 +0000 (16:30 -0700)]
thp, s390: disable thp for kvm host on s390
This patch is part of the architecture backend for thp on s390. It
disables thp for kvm hosts, because there is no kvm host hugepage support
so far. Existing thp mappings are split by follow_page() with FOLL_SPLIT,
and future thp mappings are prevented by setting VM_NOHUGEPAGE in
mm->def_flags.
Gerald Schaefer [Mon, 8 Oct 2012 23:30:18 +0000 (16:30 -0700)]
thp, s390: thp pagetable pre-allocation for s390
This patch is part of the architecture backend for thp on s390. It
provides the pagetable pre-allocation functions
pgtable_trans_huge_deposit() and pgtable_trans_huge_withdraw(). Unlike
other archs, s390 has no struct page * as pgtable_t, but rather a pointer
to the page table. So instead of saving the pagetable pre- allocation
list info inside the struct page, it is being saved within the pagetable
itself.
Gerald Schaefer [Mon, 8 Oct 2012 23:30:15 +0000 (16:30 -0700)]
thp, s390: thp splitting backend for s390
This patch is part of the architecture backend for thp on s390. It
provides the functions related to thp splitting, including serialization
against gup. Unlike other archs, pmdp_splitting_flush() cannot use a tlb
flushing operation to serialize against gup on s390, because that wouldn't
be stopped by the disabled IRQs. So instead, smp_call_function() is
called with an empty function, which will have the expected effect.
Gerald Schaefer [Mon, 8 Oct 2012 23:30:12 +0000 (16:30 -0700)]
thp: make MADV_HUGEPAGE check for mm->def_flags
This adds a check to hugepage_madvise(), to refuse MADV_HUGEPAGE if
VM_NOHUGEPAGE is set in mm->def_flags. On s390, the VM_NOHUGEPAGE flag
will be set in mm->def_flags for kvm processes, to prevent any future thp
mappings. In order to also prevent MADV_HUGEPAGE on such an mm,
hugepage_madvise() should check mm->def_flags.
Gerald Schaefer [Mon, 8 Oct 2012 23:30:09 +0000 (16:30 -0700)]
thp: introduce pmdp_invalidate()
On s390, a valid page table entry must not be changed while it is attached
to any CPU. So instead of pmd_mknotpresent() and set_pmd_at(), an IDTE
operation would be necessary there. This patch introduces the
pmdp_invalidate() function, to allow architecture-specific
implementations.
Gerald Schaefer [Mon, 8 Oct 2012 23:30:07 +0000 (16:30 -0700)]
thp: remove assumptions on pgtable_t type
The thp page table pre-allocation code currently assumes that pgtable_t is
of type "struct page *". This may not be true for all architectures, so
this patch removes that assumption by replacing the functions
prepare_pmd_huge_pte() and get_pmd_huge_pte() with two new functions that
can be defined architecture-specific.
It also removes two VM_BUG_ON checks for page_count() and page_mapcount()
operating on a pgtable_t. Apart from the VM_BUG_ON removal, there will be
no functional change introduced by this patch.
Cleanup patch in preparation for transparent hugepage support on s390.
Adding new architectures to the TRANSPARENT_HUGEPAGE config option can
make the "depends" line rather ugly, like "depends on (X86 || (S390 &&
64BIT)) && MMU".
This patch adds a HAVE_ARCH_TRANSPARENT_HUGEPAGE instead. x86 already has
MMU "def_bool y", so the MMU check is superfluous there and
HAVE_ARCH_TRANSPARENT_HUGEPAGE can be selected in arch/x86/Kconfig.
Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Hugh Dickins <hughd@google.com> Cc: Hillf Danton <dhillf@gmail.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: 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>
Xiao Guangrong [Mon, 8 Oct 2012 23:29:48 +0000 (16:29 -0700)]
thp: merge page pre-alloc in khugepaged_loop into khugepaged_do_scan
There are two pre-alloc operations in these two function, the different is:
- it allows to sleep if page alloc fail in khugepaged_loop
- it exits immediately if page alloc fail in khugepaged_do_scan
Actually, in khugepaged_do_scan, we can allow the pre-alloc to sleep on
the first failure, then the operation in khugepaged_loop can be removed
Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Xiao Guangrong [Mon, 8 Oct 2012 23:29:44 +0000 (16:29 -0700)]
thp: remove wake_up_interruptible in the exit path
Add the check of kthread_should_stop() to the conditions which are used to
wakeup on khugepaged_wait, then kthread_stop is enough to let the thread
exit
Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Xiao Guangrong [Mon, 8 Oct 2012 23:29:42 +0000 (16:29 -0700)]
thp: remove unnecessary khugepaged_thread check
Now, khugepaged creation and cancel are completely serial under the
protection of khugepaged_mutex, it is impossible that many khugepaged
entities are running
Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Xiao Guangrong [Mon, 8 Oct 2012 23:29:41 +0000 (16:29 -0700)]
thp: move khugepaged_mutex out of khugepaged
Currently, hugepaged_mutex is used really complexly and hard to
understand, actually, it is just used to serialize start_khugepaged and
khugepaged for these reasons:
- khugepaged_thread is shared between them
- the thp disable path (echo never > transparent_hugepage/enabled) is
nonblocking, so we need to protect khugepaged_thread to get a stable
running state
These can be avoided by:
- use the lock to serialize the thread creation and cancel
- thp disable path can not finised until the thread exits
Then khugepaged_thread is fully controlled by start_khugepaged, khugepaged
will be happy without the lock
Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Make sure the #endif that terminates the standard #ifndef / #define /
#endif construct gets labeled, and gets positioned at the end of the file
as is normally the case.
Signed-off-by: Michel Lespinasse <walken@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Will Deacon [Mon, 8 Oct 2012 23:29:32 +0000 (16:29 -0700)]
mm: hugetlb: add arch hook for clearing page flags before entering pool
The core page allocator ensures that page flags are zeroed when freeing
pages via free_pages_check. A number of architectures (ARM, PPC, MIPS)
rely on this property to treat new pages as dirty with respect to the data
cache and perform the appropriate flushing before mapping the pages into
userspace.
This can lead to cache synchronisation problems when using hugepages,
since the allocator keeps its own pool of pages above the usual page
allocator and does not reset the page flags when freeing a page into the
pool.
This patch adds a new architecture hook, arch_clear_hugepage_flags, so
that architectures which rely on the page flags being in a particular
state for fresh allocations can adjust the flags accordingly when a page
is freed into the pool.
Signed-off-by: Will Deacon <will.deacon@arm.com> Cc: Michal Hocko <mhocko@suse.cz> Reviewed-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Gavin Shan [Mon, 8 Oct 2012 23:29:26 +0000 (16:29 -0700)]
mm/mmu_notifier: init notifier if necessary
While registering MMU notifier, new instance of MMU notifier_mm will be
allocated and later free'd if currrent mm_struct's MMU notifier_mm has
been initialized. That causes some overhead. The patch tries to
elominate that.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com> Signed-off-by: Wanpeng Li <liwanp@linux.vnet.ibm.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Avi Kivity <avi@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Cc: Sagi Grimberg <sagig@mellanox.co.il> Cc: Haggai Eran <haggaie@mellanox.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sagi Grimberg [Mon, 8 Oct 2012 23:29:24 +0000 (16:29 -0700)]
mm: mmu_notifier: have mmu_notifiers use a global SRCU so they may safely schedule
With an RCU based mmu_notifier implementation, any callout to
mmu_notifier_invalidate_range_{start,end}() or
mmu_notifier_invalidate_page() would not be allowed to call schedule()
as that could potentially allow a modification to the mmu_notifier
structure while it is currently being used.
Since srcu allocs 4 machine words per instance per cpu, we may end up
with memory exhaustion if we use srcu per mm. So all mms share a global
srcu. Note that during large mmu_notifier activity exit & unregister
paths might hang for longer periods, but it is tolerable for current
mmu_notifier clients.
Signed-off-by: Sagi Grimberg <sagig@mellanox.co.il> Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Haggai Eran <haggaie@mellanox.com> Cc: "Paul E. McKenney" <paulmck@us.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Xiao Guangrong [Mon, 8 Oct 2012 23:29:23 +0000 (16:29 -0700)]
mm: mmu_notifier: fix inconsistent memory between secondary MMU and host
There is a bug in set_pte_at_notify() which always sets the pte to the
new page before releasing the old page in the secondary MMU. At this
time, the process will access on the new page, but the secondary MMU
still access on the old page, the memory is inconsistent between them
The below scenario shows the bug more clearly:
at the beginning: *p = 0, and p is write-protected by KSM or shared with
parent process
CPU 0 CPU 1
write 1 to p to trigger COW,
set_pte_at_notify will be called:
*pte = new_page + W; /* The W bit of pte is set */
*p = 1; /* pte is valid, so no #PF */
return back to secondary MMU, then
the secondary MMU read p, but get:
*p == 0;
/*
* !!!!!!
* the host has already set p to 1, but the secondary
* MMU still get the old value 0
*/
call mmu_notifier_change_pte to release
old page in secondary MMU
We can fix it by release old page first, then set the pte to the new
page.
Note, the new page will be firstly used in secondary MMU before it is
mapped into the page table of the process, but this is safe because it
is protected by the page table lock, there is no race to change the pte
[akpm@linux-foundation.org: add comment from Andrea] Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Cc: Avi Kivity <avi@redhat.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:29:20 +0000 (16:29 -0700)]
mempolicy: fix a memory corruption by refcount imbalance in alloc_pages_vma()
Commit cc9a6c877661 ("cpuset: mm: reduce large amounts of memory barrier
related damage v3") introduced a potential memory corruption.
shmem_alloc_page() uses a pseudo vma and it has one significant unique
combination, vma->vm_ops=NULL and vma->policy->flags & MPOL_F_SHARED.
get_vma_policy() does NOT increase a policy ref when vma->vm_ops=NULL
and mpol_cond_put() DOES decrease a policy ref when a policy has
MPOL_F_SHARED. Therefore, when a cpuset update race occurs,
alloc_pages_vma() falls in 'goto retry_cpuset' path, decrements the
reference count and frees the policy prematurely.
KOSAKI Motohiro [Mon, 8 Oct 2012 23:29:19 +0000 (16:29 -0700)]
mempolicy: fix refcount leak in mpol_set_shared_policy()
When shared_policy_replace() fails to allocate new->policy is not freed
correctly by mpol_set_shared_policy(). The problem is that shared
mempolicy code directly call kmem_cache_free() in multiple places where
it is easy to make a mistake.
This patch creates an sp_free wrapper function and uses it. The bug was
introduced pre-git age (IOW, before 2.6.12-rc2).
Mel Gorman [Mon, 8 Oct 2012 23:29:17 +0000 (16:29 -0700)]
mempolicy: fix a race in shared_policy_replace()
shared_policy_replace() use of sp_alloc() is unsafe. 1) sp_node cannot
be dereferenced if sp->lock is not held and 2) another thread can modify
sp_node between spin_unlock for allocating a new sp node and next
spin_lock. The bug was introduced before 2.6.12-rc2.
Kosaki's original patch for this problem was to allocate an sp node and
policy within shared_policy_replace and initialise it when the lock is
reacquired. I was not keen on this approach because it partially
duplicates sp_alloc(). As the paths were sp->lock is taken are not that
performance critical this patch converts sp->lock to sp->mutex so it can
sleep when calling sp_alloc().
[kosaki.motohiro@jp.fujitsu.com: Original patch] Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Reviewed-by: Christoph Lameter <cl@linux.com> Cc: Josh Boyer <jwboyer@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The problem is that the structure is being prematurely freed due to a
reference count imbalance. In the following case mbind(addr, len) should
replace the memory policies of both vma1 and vma2 and thus they will
become to share the same mempolicy and the new mempolicy will have the
MPOL_F_SHARED flag.
alloc_pages_vma() uses get_vma_policy() and mpol_cond_put() pair for
maintaining the mempolicy reference count. The current rule is that
get_vma_policy() only increments refcount for shmem VMA and
mpol_conf_put() only decrements refcount if the policy has
MPOL_F_SHARED.
In above case, vma1 is not shmem vma and vma->policy has MPOL_F_SHARED!
The reference count will be decreased even though was not increased
whenever alloc_page_vma() is called. This has been broken since commit
[52cd3b07: mempolicy: rework mempolicy Reference Counting] in 2008.
There is another serious bug with the sharing of memory policies.
Currently, mempolicy rebind logic (it is called from cpuset rebinding)
ignores a refcount of mempolicy and override it forcibly. Thus, any
mempolicy sharing may cause mempolicy corruption. The bug was
introduced by commit [68860ec1: cpusets: automatic numa mempolicy
rebinding].
Ideally, the shared policy handling would be rewritten to either
properly handle COW of the policy structures or at least reference count
MPOL_F_SHARED based exclusively on information within the policy.
However, this patch takes the easier approach of disabling any policy
sharing between VMAs. Each new range allocated with sp_alloc will
allocate a new policy, set the reference count to 1 and drop the
reference count of the old policy. This increases the memory footprint
but is not expected to be a major problem as mbind() is unlikely to be
used for fine-grained ranges. It is also inefficient because it means
we allocate a new policy even in cases where mbind_range() could use the
new_policy passed to it. However, it is more straight-forward and the
change should be invisible to the user.
[mgorman@suse.de: Edited changelog] Reported-by: Dave Jones <davej@redhat.com>, Cc: Christoph Lameter <cl@linux.com>, Reviewed-by: Christoph Lameter <cl@linux.com> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Mel Gorman <mgorman@suse.de> Cc: Josh Boyer <jwboyer@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
KOSAKI Motohiro [Mon, 8 Oct 2012 23:29:14 +0000 (16:29 -0700)]
revert "mm: mempolicy: Let vma_merge and vma_split handle vma->vm_policy linkages"
Commit 05f144a0d5c2 ("mm: mempolicy: Let vma_merge and vma_split handle
vma->vm_policy linkages") removed vma->vm_policy updates code but it is
the purpose of mbind_range(). Now, mbind_range() is virtually a no-op
and while it does not allow memory corruption it is not the right fix.
This patch is a revert.
Mel Gorman [Mon, 8 Oct 2012 23:29:12 +0000 (16:29 -0700)]
mm: compaction: capture a suitable high-order page immediately when it is made available
While compaction is migrating pages to free up large contiguous blocks
for allocation it races with other allocation requests that may steal
these blocks or break them up. This patch alters direct compaction to
capture a suitable free page as soon as it becomes available to reduce
this race. It uses similar logic to split_free_page() to ensure that
watermarks are still obeyed.
Signed-off-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Rik van Riel <riel@redhat.com> Reviewed-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:29:11 +0000 (16:29 -0700)]
mm: vmscan: scale number of pages reclaimed by reclaim/compaction based on failures
If allocation fails after compaction then compaction may be deferred for
a number of allocation attempts. If there are subsequent failures,
compact_defer_shift is increased to defer for longer periods. This
patch uses that information to scale the number of pages reclaimed with
compact_defer_shift until allocations succeed again. The rationale is
that reclaiming the normal number of pages still allowed compaction to
fail and its success depends on the number of pages. If it's failing,
reclaim more pages until it succeeds again.
Note that this is not implying that VM reclaim is not reclaiming enough
pages or that its logic is broken. try_to_free_pages() always asks for
SWAP_CLUSTER_MAX pages to be reclaimed regardless of order and that is
what it does. Direct reclaim stops normally with this check.
if (sc->nr_reclaimed >= sc->nr_to_reclaim)
goto out;
should_continue_reclaim delays when that check is made until a minimum
number of pages for reclaim/compaction are reclaimed. It is possible
that this patch could instead set nr_to_reclaim in try_to_free_pages()
and drive it from there but that's behaves differently and not
necessarily for the better. If driven from do_try_to_free_pages(), it
is also possible that priorities will rise.
When they reach DEF_PRIORITY-2, it will also start stalling and setting
pages for immediate reclaim which is more disruptive than not desirable
in this case. That is a more wide-reaching change that could cause
another regression related to THP requests causing interactive jitter.
[akpm@linux-foundation.org: fix build] Signed-off-by: Mel Gorman <mgorman@suse.de> Acked-by: Rik van Riel <riel@redhat.com> Reviewed-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Mon, 8 Oct 2012 23:29:09 +0000 (16:29 -0700)]
mm: compaction: update comment in try_to_compact_pages
Allocation success rates have been far lower since 3.4 due to commit fe2c2a106663 ("vmscan: reclaim at order 0 when compaction is enabled").
This commit was introduced for good reasons and it was known in advance
that the success rates would suffer but it was justified on the grounds
that the high allocation success rates were achieved by aggressive
reclaim. Success rates are expected to suffer even more in 3.6 due to
commit 7db8889ab05b ("mm: have order > 0 compaction start off where it
left") which testing has shown to severely reduce allocation success
rates under load - to 0% in one case.
This series aims to improve the allocation success rates without
regressing the benefits of commit fe2c2a106663. The series is based on
latest mmotm and takes into account the __GFP_NO_KSWAPD flag is going
away.
Patch 1 updates a stale comment seeing as I was in the general area.
Patch 2 updates reclaim/compaction to reclaim pages scaled on the number
of recent failures.
Patch 3 captures suitable high-order pages freed by compaction to reduce
races with parallel allocation requests.
Patch 4 fixes the upstream commit [7db8889a: mm: have order > 0 compaction
start off where it left] to enable compaction again
Patch 5 identifies when compacion is taking too long due to contention
and aborts.
I know that the allocation success rates in 3.3.6 was 78% in comparison
to 36% in in the current akpm tree. With the full series applied, the
success rates are up to around 51% with some variability in the results.
This is not as high a success rate but it does not reclaim excessively
which is a key point.
Note that swap in/out rates remain at 0. In 3.3.6 with 78% success rates
there were 71881 pages swapped out.
Direct pages scanned 70942 122976
Kswapd pages scanned 13663001520122
Kswapd pages reclaimed 13662141484629
Direct pages reclaimed 70936 105716
Kswapd efficiency 99% 97%
Kswapd velocity 1072.550 1182.615
Direct efficiency 99% 85%
Direct velocity 55.690 95.672
The kswapd velocity changes very little as expected. kswapd velocity is
around the 1000 pages/sec mark where as in kernel 3.3.6 with the high
allocation success rates it was 8140 pages/second. Direct velocity is
higher as a result of patch 2 of the series but this is expected and is
acceptable. The direct reclaim and kswapd velocities change very little.
If these get accepted for merging then there is a difficulty in how they
should be handled. 7db8889a ("mm: have order > 0 compaction start off
where it left") is broken but it is already in 3.6-rc1 and needs to be
fixed. However, if just patch 4 from this series is applied then Jim
Schutt's workload is known to break again as his workload also requires
patch 5. While it would be preferred to have all these patches in 3.6 to
improve compaction in general, it would at least be acceptable if just
patches 4 and 5 were merged to 3.6 to fix a known problem without breaking
compaction completely. On the face of it, that would force
__GFP_NO_KSWAPD patches to be merged at the same time but I can do a
version of this series with __GFP_NO_KSWAPD change reverted and then
rebase it on top of this series. That might be best overall because I
note that the __GFP_NO_KSWAPD patch should have removed
deferred_compaction from page_alloc.c but it didn't but fixing that causes
collisions with this series.
This patch:
The comment about order applied when the check was order >
PAGE_ALLOC_COSTLY_ORDER which has not been the case since c5a73c3d ("thp:
use compaction for all allocation orders"). Fixing the comment while I'm
in the general area.
Signed-off-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Rik van Riel <riel@redhat.com> Reviewed-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
projects / mirror_ubuntu-eoan-kernel.git / log