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>
SLUB: Avoid page struct cacheline bouncing due to remote frees to cpu slab
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>
Mel Gorman [Tue, 16 Oct 2007 08:26:02 +0000 (01:26 -0700)]
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> Acked-by: Andy Whitcroft <apw@shadowen.org> 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>
Mel Gorman [Tue, 16 Oct 2007 08:26:01 +0000 (01:26 -0700)]
Do not depend on MAX_ORDER when grouping pages by mobility
Currently mobility grouping works at the MAX_ORDER_NR_PAGES level. This makes
sense for the majority of users where this is also the huge page size.
However, on platforms like ia64 where the huge page size is runtime
configurable it is desirable to group at a lower order. On x86_64 and
occasionally on x86, the hugepage size may not always be MAX_ORDER_NR_PAGES.
This patch groups pages together based on the value of HUGETLB_PAGE_ORDER. It
uses a compile-time constant if possible and a variable where the huge page
size is runtime configurable.
It is assumed that grouping should be done at the lowest sensible order and
that the user would not want to override this. If this is not true,
page_block order could be forced to a variable initialised via a boot-time
kernel parameter.
One potential issue with this patch is that IA64 now parses hugepagesz with
early_param() instead of __setup(). __setup() is called after the memory
allocator has been initialised and the pageblock bitmaps already setup. In
tests on one IA64 there did not seem to be any problem with using
early_param() and in fact may be more correct as it guarantees the parameter
is handled before the parsing of hugepages=.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> Acked-by: Andy Whitcroft <apw@shadowen.org> 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>
Mel Gorman [Tue, 16 Oct 2007 08:26:00 +0000 (01:26 -0700)]
Fix calculation in move_freepages_block for counting pages
move_freepages_block() returns the number of blocks moved. This value is used
to determine if a block of pages should be stolen for the exclusive use of a
migrate type or not. However, the value returned is being used correctly.
This patch fixes the calculation to return the number of base pages that have
been moved.
This should be considered a fix to the patch
move-free-pages-between-lists-on-steal.patch
Credit to Andy Whitcroft for spotting the problem.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> Acked-by: Andy Whitcroft <apw@shadowen.org> 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>
Mel Gorman [Tue, 16 Oct 2007 08:25:59 +0000 (01:25 -0700)]
don't group high order atomic allocations
Grouping high-order atomic allocations together was intended to allow
bursty users of atomic allocations to work such as e1000 in situations
where their preallocated buffers were depleted. This did not work in at
least one case with a wireless network adapter needing order-1 allocations
frequently. To resolve that, the free pages used for min_free_kbytes were
moved to separate contiguous blocks with the patch
bias-the-location-of-pages-freed-for-min_free_kbytes-in-the-same-max_order_nr_pages-blocks.
It is felt that keeping the free pages in the same contiguous blocks should
be sufficient for bursty short-lived high-order atomic allocations to
succeed, maybe even with the e1000. Even if there is a failure, increasing
the value of min_free_kbytes will free pages as contiguous bloks in
contrast to the standard buddy allocator which makes no attempt to keep the
minimum number of free pages contiguous.
This patch backs out grouping high order atomic allocations together to
determine if it is really needed or not. If a new report comes in about
high-order atomic allocations failing, the feature can be reintroduced to
determine if it fixes the problem or not. As a side-effect, this patch
reduces by 1 the number of bits required to track the mobility type of
pages within a MAX_ORDER_NR_PAGES block.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> 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>
Mel Gorman [Tue, 16 Oct 2007 08:25:58 +0000 (01:25 -0700)]
remove PAGE_GROUP_BY_MOBILITY
Grouping pages by mobility can be disabled at compile-time. This was
considered undesirable by a number of people. However, in the current stack of
patches, it is not a simple case of just dropping the configurable patch as it
would cause merge conflicts. This patch backs out the configuration option.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> 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>
Mel Gorman [Tue, 16 Oct 2007 08:25:58 +0000 (01:25 -0700)]
Bias the location of pages freed for min_free_kbytes in the same MAX_ORDER_NR_PAGES blocks
The standard buddy allocator always favours the smallest block of pages.
The effect of this is that the pages free to satisfy min_free_kbytes tends
to be preserved since boot time at the same location of memory ffor a very
long time and as a contiguous block. When an administrator sets the
reserve at 16384 at boot time, it tends to be the same MAX_ORDER blocks
that remain free. This allows the occasional high atomic allocation to
succeed up until the point the blocks are split. In practice, it is
difficult to split these blocks but when they do split, the benefit of
having min_free_kbytes for contiguous blocks disappears. Additionally,
increasing min_free_kbytes once the system has been running for some time
has no guarantee of creating contiguous blocks.
On the other hand, CONFIG_PAGE_GROUP_BY_MOBILITY favours splitting large
blocks when there are no free pages of the appropriate type available. A
side-effect of this is that all blocks in memory tends to be used up and
the contiguous free blocks from boot time are not preserved like in the
vanilla allocator. This can cause a problem if a new caller is unwilling
to reclaim or does not reclaim for long enough.
A failure scenario was found for a wireless network device allocating
order-1 atomic allocations but the allocations were not intense or frequent
enough for a whole block of pages to be preserved for MIGRATE_HIGHALLOC.
This was reproduced on a desktop by booting with mem=256mb, forcing the
driver to allocate at order-1, running a bittorrent client (downloading a
debian ISO) and building a kernel with -j2.
This patch addresses the problem on the desktop machine booted with
mem=256mb. It works by setting aside a reserve of MAX_ORDER_NR_PAGES
blocks, the number of which depends on the value of min_free_kbytes. These
blocks are only fallen back to when there is no other free pages. Then the
smallest possible page is used just like the normal buddy allocator instead
of the largest possible page to preserve contiguous pages The pages in free
lists in the reserve blocks are never taken for another migrate type. The
results is that even if min_free_kbytes is set to a low value, contiguous
blocks will be preserved in the MIGRATE_RESERVE blocks.
This works better than the vanilla allocator because if min_free_kbytes is
increased, a new reserve block will be chosen based on the location of
reclaimable pages and the block will free up as contiguous pages. In the
vanilla allocator, no effort is made to target a block of pages to free as
contiguous pages and min_free_kbytes pages are scattered randomly.
This effect has been observed on the test machine. min_free_kbytes was set
initially low but it was kept as a contiguous free block within
MIGRATE_RESERVE. min_free_kbytes was then set to a higher value and over a
period of time, the free blocks were within the reserve and coalescing.
How long it takes to free up depends on how quickly LRU is rotating.
Amusingly, this means that more activity will free the blocks faster.
This mechanism potentially replaces MIGRATE_HIGHALLOC as it may be more
effective than grouping contiguous free pages together. It all depends on
whether the number of active atomic high allocations exceeds
min_free_kbytes or not. If the number of active allocations exceeds
min_free_kbytes, it's worth it but maybe in that situation, min_free_kbytes
should be set higher. Once there are no more reports of allocation
failures, a patch will be submitted that backs out MIGRATE_HIGHALLOC and
see if the reports stay missing.
Credit to Mariusz Kozlowski for discovering the problem, describing the
failure scenario and testing patches and scenarios.
[akpm@linux-foundation.org: cleanups] Signed-off-by: Mel Gorman <mel@csn.ul.ie> 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>
Mel Gorman [Tue, 16 Oct 2007 08:25:56 +0000 (01:25 -0700)]
Fix corruption of memmap on IA64 SPARSEMEM when mem_section is not a power of 2
There are problems in the use of SPARSEMEM and pageblock flags that causes
problems on ia64.
The first part of the problem is that units are incorrect in
SECTION_BLOCKFLAGS_BITS computation. This results in a map_section's
section_mem_map being treated as part of a bitmap which isn't good. This
was evident with an invalid virtual address when mem_init attempted to free
bootmem pages while relinquishing control from the bootmem allocator.
The second part of the problem occurs because the pageblock flags bitmap is
be located with the mem_section. The SECTIONS_PER_ROOT computation using
sizeof (mem_section) may not be a power of 2 depending on the size of the
bitmap. This renders masks and other such things not power of 2 base.
This issue was seen with SPARSEMEM_EXTREME on ia64. This patch moves the
bitmap outside of mem_section and uses a pointer instead in the
mem_section. The bitmaps are allocated when the section is being
initialised.
Note that sparse_early_usemap_alloc() does not use alloc_remap() like
sparse_early_mem_map_alloc(). The allocation required for the bitmap on
x86, the only architecture that uses alloc_remap is typically smaller than
a cache line. alloc_remap() pads out allocations to the cache size which
would be a needless waste.
Credit to Bob Picco for identifying the original problem and effecting a
fix for the SECTION_BLOCKFLAGS_BITS calculation. Credit to Andy Whitcroft
for devising the best way of allocating the bitmaps only when required for
the section.
[wli@holomorphy.com: warning fix] Signed-off-by: Bob Picco <bob.picco@hp.com> Signed-off-by: Andy Whitcroft <apw@shadowen.org> Signed-off-by: Mel Gorman <mel@csn.ul.ie> Cc: "Luck, Tony" <tony.luck@intel.com> Signed-off-by: William Irwin <bill.irwin@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Tue, 16 Oct 2007 08:25:55 +0000 (01:25 -0700)]
Be more agressive about stealing when MIGRATE_RECLAIMABLE allocations fallback
MIGRATE_RECLAIMABLE allocations tend to be very bursty in nature like when
updatedb starts. It is likely this will occur in situations where MAX_ORDER
blocks of pages are not free. This means that updatedb can scatter
MIGRATE_RECLAIMABLE pages throughout the address space. This patch is more
agressive about stealing blocks of pages for MIGRATE_RECLAIMABLE.
Mel Gorman [Tue, 16 Oct 2007 08:25:54 +0000 (01:25 -0700)]
Bias the placement of kernel pages at lower PFNs
This patch chooses blocks with lower PFNs when placing kernel allocations.
This is particularly important during fallback in low memory situations to
stop unmovable pages being placed throughout the entire address space.
Mel Gorman [Tue, 16 Oct 2007 08:25:54 +0000 (01:25 -0700)]
Do not group pages by mobility type on low memory systems
Grouping pages by mobility can only successfully operate when there are more
MAX_ORDER_NR_PAGES areas than mobility types. When there are insufficient
areas, fallbacks cannot be avoided. This has noticeable performance impacts
on machines with small amounts of memory in comparison to MAX_ORDER_NR_PAGES.
For example, on IA64 with a configuration including huge pages spans 1GiB with
MAX_ORDER_NR_PAGES so would need at least 4GiB of RAM before grouping pages by
mobility would be useful. In comparison, an x86 would need 16MB.
This patch checks the size of vm_total_pages in build_all_zonelists(). If
there are not enough areas, mobility is effectivly disabled by considering
all allocations as the same type (UNMOVABLE). This is achived via a
__read_mostly flag.
With this patch, performance is comparable to disabling grouping pages
by mobility at compile-time on a test machine with insufficient memory.
With this patch, it is reasonable to get rid of grouping pages by mobility
a compile-time option.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> 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>
Mel Gorman [Tue, 16 Oct 2007 08:25:53 +0000 (01:25 -0700)]
Group high-order atomic allocations
In rare cases, the kernel needs to allocate a high-order block of pages
without sleeping. For example, this is the case with e1000 cards configured
to use jumbo frames. Migrating or reclaiming pages in this situation is not
an option.
This patch groups these allocations together as much as possible by adding a
new MIGRATE_TYPE. The MIGRATE_HIGHATOMIC type are exactly what they sound
like. Care is taken that pages of other migrate types do not use the same
blocks as high-order atomic allocations.
Mel Gorman [Tue, 16 Oct 2007 08:25:52 +0000 (01:25 -0700)]
Group short-lived and reclaimable kernel allocations
This patch marks a number of allocations that are either short-lived such as
network buffers or are reclaimable such as inode allocations. When something
like updatedb is called, long-lived and unmovable kernel allocations tend to
be spread throughout the address space which increases fragmentation.
This patch groups these allocations together as much as possible by adding a
new MIGRATE_TYPE. The MIGRATE_RECLAIMABLE type is for allocations that can be
reclaimed on demand, but not moved. i.e. they can be migrated by deleting
them and re-reading the information from elsewhere.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> Cc: Andy Whitcroft <apw@shadowen.org> Cc: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Tue, 16 Oct 2007 08:25:51 +0000 (01:25 -0700)]
Move free pages between lists on steal
When a fallback occurs, there will be free pages for one allocation type
stored on the list for another. When a large steal occurs, this patch will
move all the free pages within one list to the other.
[y-goto@jp.fujitsu.com: fix BUG_ON check at move_freepages()]
[apw@shadowen.org: Move to using pfn_valid_within()] Signed-off-by: Mel Gorman <mel@csn.ul.ie> Cc: Christoph Lameter <clameter@engr.sgi.com> Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com> Cc: Bjorn Helgaas <bjorn.helgaas@hp.com> Signed-off-by: Andy Whitcroft <andyw@uk.ibm.com> Cc: Bob Picco <bob.picco@hp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Tue, 16 Oct 2007 08:25:50 +0000 (01:25 -0700)]
Drain per-cpu lists when high-order allocations fail
Per-cpu pages can accidentally cause fragmentation because they are free, but
pinned pages in an otherwise contiguous block. When this patch is applied,
the per-cpu caches are drained after the direct-reclaim is entered if the
requested order is greater than 0. It simply reuses the code used by suspend
and hotplug.
Mel Gorman [Tue, 16 Oct 2007 08:25:50 +0000 (01:25 -0700)]
Add a configure option to group pages by mobility
The grouping mechanism has some memory overhead and a more complex allocation
path. This patch allows the strategy to be disabled for small memory systems
or if it is known the workload is suffering because of the strategy. It also
acts to show where the page groupings strategy interacts with the standard
buddy allocator.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> Signed-off-by: Joel Schopp <jschopp@austin.ibm.com> Cc: Andy Whitcroft <apw@shadowen.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman [Tue, 16 Oct 2007 08:25:49 +0000 (01:25 -0700)]
Choose pages from the per-cpu list based on migration type
The freelists for each migrate type can slowly become polluted due to the
per-cpu list. Consider what happens when the following happens
1. A 2^(MAX_ORDER-1) list is reserved for __GFP_MOVABLE pages
2. An order-0 page is allocated from the newly reserved block
3. The page is freed and placed on the per-cpu list
4. alloc_page() is called with GFP_KERNEL as the gfp_mask
5. The per-cpu list is used to satisfy the allocation
This results in a kernel page is in the middle of a migratable region. This
patch prevents this leak occuring by storing the MIGRATE_ type of the page in
page->private. On allocate, a page will only be returned of the desired type,
else more pages will be allocated. This may temporarily allow a per-cpu list
to go over the pcp->high limit but it'll be corrected on the next free. Care
is taken to preserve the hotness of pages recently freed.
The additional code is not measurably slower for the workloads we've tested.
Mel Gorman [Tue, 16 Oct 2007 08:25:48 +0000 (01:25 -0700)]
Split the free lists for movable and unmovable allocations
This patch adds the core of the fragmentation reduction strategy. It works by
grouping pages together based on their ability to migrate or be reclaimed.
Basically, it works by breaking the list in zone->free_area list into
MIGRATE_TYPES number of lists.
Mel Gorman [Tue, 16 Oct 2007 08:25:47 +0000 (01:25 -0700)]
Add a bitmap that is used to track flags affecting a block of pages
Here is the latest revision of the anti-fragmentation patches. Of particular
note in this version is special treatment of high-order atomic allocations.
Care is taken to group them together and avoid grouping pages of other types
near them. Artifical tests imply that it works. I'm trying to get the
hardware together that would allow setting up of a "real" test. If anyone
already has a setup and test that can trigger the atomic-allocation problem,
I'd appreciate a test of these patches and a report. The second major change
is that these patches will apply cleanly with patches that implement
anti-fragmentation through zones.
kernbench shows effectively no performance difference varying between -0.2%
and +2% on a variety of test machines. Success rates for huge page allocation
are dramatically increased. For example, on a ppc64 machine, the vanilla
kernel was only able to allocate 1% of memory as a hugepage and this was due
to a single hugepage reserved as min_free_kbytes. With these patches applied,
17% was allocatable as superpages. With reclaim-related fixes from Andy
Whitcroft, it was 40% and further reclaim-related improvements should increase
this further.
Changelog Since V28
o Group high-order atomic allocations together
o It is no longer required to set min_free_kbytes to 10% of memory. A value
of 16384 in most cases will be sufficient
o Now applied with zone-based anti-fragmentation
o Fix incorrect VM_BUG_ON within buffered_rmqueue()
o Reorder the stack so later patches do not back out work from earlier patches
o Fix bug were journal pages were being treated as movable
o Bias placement of non-movable pages to lower PFNs
o More agressive clustering of reclaimable pages in reactions to workloads
like updatedb that flood the size of inode caches
Changelog Since V27
o Renamed anti-fragmentation to Page Clustering. Anti-fragmentation was giving
the mistaken impression that it was the 100% solution for high order
allocations. Instead, it greatly increases the chances high-order
allocations will succeed and lays the foundation for defragmentation and
memory hot-remove to work properly
o Redefine page groupings based on ability to migrate or reclaim instead of
basing on reclaimability alone
o Get rid of spurious inits
o Per-cpu lists are no longer split up per-type. Instead the per-cpu list is
searched for a page of the appropriate type
o Added more explanation commentary
o Fix up bug in pageblock code where bitmap was used before being initalised
Changelog Since V26
o Fix double init of lists in setup_pageset
Changelog Since V25
o Fix loop order of for_each_rclmtype_order so that order of loop matches args
o gfpflags_to_rclmtype uses gfp_t instead of unsigned long
o Rename get_pageblock_type() to get_page_rclmtype()
o Fix alignment problem in move_freepages()
o Add mechanism for assigning flags to blocks of pages instead of page->flags
o On fallback, do not examine the preferred list of free pages a second time
The purpose of these patches is to reduce external fragmentation by grouping
pages of related types together. When pages are migrated (or reclaimed under
memory pressure), large contiguous pages will be freed.
This patch works by categorising allocations by their ability to migrate;
Movable - The pages may be moved with the page migration mechanism. These are
generally userspace pages.
Reclaimable - These are allocations for some kernel caches that are
reclaimable or allocations that are known to be very short-lived.
Unmovable - These are pages that are allocated by the kernel that
are not trivially reclaimed. For example, the memory allocated for a
loaded module would be in this category. By default, allocations are
considered to be of this type
HighAtomic - These are high-order allocations belonging to callers that
cannot sleep or perform any IO. In practice, this is restricted to
jumbo frame allocation for network receive. It is assumed that the
allocations are short-lived
Instead of having one MAX_ORDER-sized array of free lists in struct free_area,
there is one for each type of reclaimability. Once a 2^MAX_ORDER block of
pages is split for a type of allocation, it is added to the free-lists for
that type, in effect reserving it. Hence, over time, pages of the different
types can be clustered together.
When the preferred freelists are expired, the largest possible block is taken
from an alternative list. Buddies that are split from that large block are
placed on the preferred allocation-type freelists to mitigate fragmentation.
This implementation gives best-effort for low fragmentation in all zones.
Ideally, min_free_kbytes needs to be set to a value equal to 4 * (1 <<
(MAX_ORDER-1)) pages in most cases. This would be 16384 on x86 and x86_64 for
example.
Our tests show that about 60-70% of physical memory can be allocated on a
desktop after a few days uptime. In benchmarks and stress tests, we are
finding that 80% of memory is available as contiguous blocks at the end of the
test. To compare, a standard kernel was getting < 1% of memory as large pages
on a desktop and about 8-12% of memory as large pages at the end of stress
tests.
Following this email are 12 patches that implement thie page grouping feature.
The first patch introduces a mechanism for storing flags related to a whole
block of pages. Then allocations are split between movable and all other
allocations. Following that are patches to deal with per-cpu pages and make
the mechanism configurable. The next patch moves free pages between lists
when partially allocated blocks are used for pages of another migrate type.
The second last patch groups reclaimable kernel allocations such as inode
caches together. The final patch related to groupings keeps high-order atomic
allocations.
The last two patches are more concerned with control of fragmentation. The
second last patch biases placement of non-movable allocations towards the
start of memory. This is with a view of supporting memory hot-remove of DIMMs
with higher PFNs in the future. The biasing could be enforced a lot heavier
but it would cost. The last patch agressively clusters reclaimable pages like
inode caches together.
The fragmentation reduction strategy needs to track if pages within a block
can be moved or reclaimed so that pages are freed to the appropriate list.
This patch adds a bitmap for flags affecting a whole a MAX_ORDER block of
pages.
In non-SPARSEMEM configurations, the bitmap is stored in the struct zone and
allocated during initialisation. SPARSEMEM statically allocates the bitmap in
a struct mem_section so that bitmaps do not have to be resized during memory
hotadd. This wastes a small amount of memory per unused section (usually
sizeof(unsigned long)) but the complexity of dynamically allocating the memory
is quite high.
Additional credit to Andy Whitcroft who reviewed up an earlier implementation
of the mechanism an suggested how to make it a *lot* cleaner.
Signed-off-by: Mel Gorman <mel@csn.ul.ie> Cc: Andy Whitcroft <apw@shadowen.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
flush icache before set_pte() on ia64: flush icache at set_pte
Current ia64 kernel flushes icache by lazy_mmu_prot_update() *after*
set_pte(). This is too late. This patch removes lazy_mmu_prot_update and
add modfied set_pte() for flushing if necessary.
This patch flush icache of a page when
new pte has exec bit.
&& new pte has present bit
&& new pte is user's page.
&& (old *ptep is not present
|| new pte's pfn is not same to old *ptep's ptn)
&& new pte's page has no Pg_arch_1 bit.
Pg_arch_1 is set when a page is cache consistent.
I think this condition checks are much easier to understand than considering
"Where sync_icache_dcache() should be inserted ?".
pte_user() for ia64 was removed by http://lkml.org/lkml/2007/6/12/67 as
clean-up. So, I added it again.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Luck, Tony" <tony.luck@intel.com> Cc: Christoph Lameter <clameter@sgi.com> Cc: Hugh Dickins <hugh@veritas.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Andrea Arcangeli [Tue, 16 Oct 2007 08:25:42 +0000 (01:25 -0700)]
make swappiness safer to use
Swappiness isn't a safe sysctl. Setting it to 0 for example can hang a
system. That's a corner case but even setting it to 10 or lower can waste
enormous amounts of cpu without making much progress. We've customers who
wants to use swappiness but they can't because of the current
implementation (if you change it so the system stops swapping it really
stops swapping and nothing works sane anymore if you really had to swap
something to make progress).
This patch from Kurt Garloff makes swappiness safer to use (no more huge
cpu usage or hangs with low swappiness values).
I think the prev_priority can also be nuked since it wastes 4 bytes per
zone (that would be an incremental patch but I wait the nr_scan_[in]active
to be nuked first for similar reasons). Clearly somebody at some point
noticed how broken that thing was and they had to add min(priority,
prev_priority) to give it some reliability, but they didn't go the last
mile to nuke prev_priority too. Calculating distress only in function of
not-racy priority is correct and sure more than enough without having to
add randomness into the equation.
Patch is tested on older kernels but it compiles and it's quite simple
so...
Overall I'm not very satisified by the swappiness tweak, since it doesn't
rally do anything with the dirty pagecache that may be inactive. We need
another kind of tweak that controls the inactive scan and tunes the
can_writepage feature (not yet in mainline despite having submitted it a
few times), not only the active one. That new tweak will tell the kernel
how hard to scan the inactive list for pure clean pagecache (something the
mainline kernel isn't capable of yet). We already have that feature
working in all our enterprise kernels with the default reasonable tune, or
they can't even run a readonly backup with tar without triggering huge
write I/O. I think it should be available also in mainline later.
Cc: Nick Piggin <npiggin@suse.de> Signed-off-by: Kurt Garloff <garloff@suse.de> Signed-off-by: Andrea Arcangeli <andrea@suse.de> 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>
The function of GFP_LEVEL_MASK seems to be unclear. In order to clear up
the mystery we get rid of it and replace GFP_LEVEL_MASK with 3 sets of GFP
flags:
GFP_RECLAIM_MASK Flags used to control page allocator reclaim behavior.
GFP_CONSTRAINT_MASK Flags used to limit where allocations can occur.
GFP_SLAB_BUG_MASK Flags that the slab allocator BUG()s on.
These replace the uses of GFP_LEVEL mask in the slab allocators and in
vmalloc.c.
The use of the flags not included in these sets may occur as a result of a
slab allocation standing in for a page allocation when constructing scatter
gather lists. Extraneous flags are cleared and not passed through to the
page allocator. __GFP_MOVABLE/RECLAIMABLE, __GFP_COLD and __GFP_COMP will
now be ignored if passed to a slab allocator.
Change the allocation of allocator meta data in SLAB and vmalloc to not
pass through flags listed in GFP_CONSTRAINT_MASK. SLAB already removes the
__GFP_THISNODE flag for such allocations. Generalize that to also cover
vmalloc. The use of GFP_CONSTRAINT_MASK also includes __GFP_HARDWALL.
The impact of allocator metadata placement on access latency to the
cachelines of the object itself is minimal since metadata is only
referenced on alloc and free. The attempt is still made to place the meta
data optimally but we consistently allow fallback both in SLAB and vmalloc
(SLUB does not need to allocate metadata like that).
Allocator metadata may serve multiple in kernel users and thus should not
be subject to the limitations arising from a single allocation context.
[akpm@linux-foundation.org: fix fallback_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>
Yasunori Goto [Tue, 16 Oct 2007 08:25:40 +0000 (01:25 -0700)]
Fix panic of cpu online with memory less node
When a cpu is onlined on memory-less-node box, kernel panics due to touch
NULL pointer of pgdat->kswapd. Current kswapd runs only nodes which have
memory. So, calling of set_cpus_allowed() is not necessary for memory-less
node.
Lee Schermerhorn [Tue, 16 Oct 2007 08:25:39 +0000 (01:25 -0700)]
memoryless nodes: fixup uses of node_online_map in generic code
Here's a cut at fixing up uses of the online node map in generic code.
mm/shmem.c:shmem_parse_mpol()
Ensure nodelist is subset of nodes with memory.
Use node_states[N_HIGH_MEMORY] as default for missing
nodelist for interleave policy.
mm/shmem.c:shmem_fill_super()
initialize policy_nodes to node_states[N_HIGH_MEMORY]
mm/page-writeback.c:highmem_dirtyable_memory()
sum over nodes with memory
mm/page_alloc.c:zlc_setup()
allowednodes - use nodes with memory.
mm/page_alloc.c:default_zonelist_order()
average over nodes with memory.
mm/page_alloc.c:find_next_best_node()
skip nodes w/o memory.
N_HIGH_MEMORY state mask may not be initialized at this time,
unless we want to depend on early_calculate_totalpages() [see
below]. Will ZONE_MOVABLE ever be configurable?
This required calling early_calculate_totalpages()
unconditionally, and populating N_HIGH_MEMORY node
state therein from nodes in the early_node_map[].
If we can depend on this, we can eliminate the
population of N_HIGH_MEMORY mask from __build_all_zonelists()
and use the N_HIGH_MEMORY mask in find_next_best_node().
mm/mempolicy.c:mpol_check_policy()
Ensure nodes specified for policy are subset of
nodes with memory.
[akpm@linux-foundation.org: fix warnings] Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Christoph Lameter <clameter@sgi.com> Cc: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
cpusets try to ensure that any node added to a cpuset's mems_allowed is
on-line and contains memory. The assumption was that online nodes contained
memory. Thus, it is possible to add memoryless nodes to a cpuset and then add
tasks to this cpuset. This results in continuous series of oom-kill and
apparent system hang.
Change cpusets to use node_states[N_HIGH_MEMORY] [a.k.a. node_memory_map] in
place of node_online_map when vetting memories. Return error if admin
attempts to write a non-empty mems_allowed node mask containing only
memoryless-nodes.
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Signed-off-by: Bob Picco <bob.picco@hp.com> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> 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>
GFP_THISNODE checks that the zone selected is within the pgdat (node) of the
first zone of a nodelist. That only works if the node has memory. A
memoryless node will have its first node on another pgdat (node).
GFP_THISNODE currently will return simply memory on the first pgdat. Thus it
is returning memory on other nodes. GFP_THISNODE should fail if there is no
local memory on a node.
Add a new set of zonelists for each node that only contain the nodes that
belong to the zones itself so that no fallback is possible.
Then modify gfp_type to pickup the right zone based on the presence of
__GFP_THISNODE.
Drop the existing GFP_THISNODE checks from the page_allocators hot path.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Acked-by: Nishanth Aravamudan <nacc@us.ibm.com> Tested-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Memoryless nodes: drop one memoryless node boot warning
get_pfn_range_for_nid() is called multiple times for each node at boot time.
Each time, it will warn about nodes with no memory, resulting in boot messages
like:
Node 0 active with no memory
Node 0 active with no memory
Node 0 active with no memory
Node 0 active with no memory
Node 0 active with no memory
Node 0 active with no memory
On node 0 totalpages: 0
Node 0 active with no memory
Node 0 active with no memory
DMA zone: 0 pages used for memmap
Node 0 active with no memory
Node 0 active with no memory
Normal zone: 0 pages used for memmap
Node 0 active with no memory
Node 0 active with no memory
Movable zone: 0 pages used for memmap
and so on for each memoryless node.
We already have the "On node N totalpages: ..." and other related messages, so
drop the "Node N active with no memory" warnings.
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Cc: Bob Picco <bob.picco@hp.com> Cc: Nishanth Aravamudan <nacc@us.ibm.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> 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>
Memoryless nodes: Update memory policy and page migration
Online nodes now may have no memory. The checks and initialization must
therefore be changed to no longer use the online functions.
This will correctly initialize the interleave on bootup to only target nodes
with memory and will make sys_move_pages return an error when a page is to be
moved to a memoryless node. Similarly we will get an error if MPOL_BIND and
MPOL_INTERLEAVE is used on a memoryless node.
These are somewhat new semantics. So far one could specify memoryless nodes
and we would maybe do the right thing and just ignore the node (or we'd do
something strange like with MPOL_INTERLEAVE). If we want to allow the
specification of memoryless nodes via memory policies then we need to keep
checking for online nodes.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Acked-by: Nishanth Aravamudan <nacc@us.ibm.com> Tested-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Memoryless nodes: Allow profiling data to fall back to other nodes
Processors on memoryless nodes must be able to fall back to remote nodes in
order to get a profiling buffer. This may lead to excessive NUMA traffic but
I think we should allow this rather than failing.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Acked-by: Nishanth Aravamudan <nacc@us.ibm.com> Acked-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The checks for node_online in the uncached allocator are made to make sure
that memory is available on these nodes. Thus switch all the checks to use
N_HIGH_MEMORY and to N_ONLINE.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Acked-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: Nishanth Aravamudan <nacc@us.ibm.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Simply switch all for_each_online_node to for_each_node_state(NORMAL_MEMORY).
That way SLUB only operates on nodes with regular memory. Any allocation
attempt on a memoryless node or a node with just highmem will fall whereupon
SLUB will fetch memory from a nearby node (depending on how memory policies
and cpuset describe fallback).
Signed-off-by: Christoph Lameter <clameter@sgi.com> Tested-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: Nishanth Aravamudan <nacc@us.ibm.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Slab should not allocate control structures for nodes without memory. This
may seem to work right now but its unreliable since not all allocations can
fall back due to the use of GFP_THISNODE.
Switching a few for_each_online_node's to N_NORMAL_MEMORY will allow us to
only allocate for nodes that have regular memory.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Acked-by: Nishanth Aravamudan <nacc@us.ibm.com> Acked-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Memoryless nodes: Fix interleave behavior for memoryless nodes
MPOL_INTERLEAVE currently simply loops over all nodes. Allocations on
memoryless nodes will be redirected to nodes with memory. This results in an
imbalance because the neighboring nodes to memoryless nodes will get
significantly more interleave hits that the rest of the nodes on the system.
We can avoid this imbalance by clearing the nodes in the interleave node set
that have no memory. If we use the node map of the memory nodes instead of
the online nodes then we have only the nodes we want.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Tested-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Memoryless nodes: introduce mask of nodes with memory
It is necessary to know if nodes have memory since we have recently begun to
add support for memoryless nodes. For that purpose we introduce a two new
node states: N_HIGH_MEMORY and N_NORMAL_MEMORY.
A node has its bit in N_HIGH_MEMORY set if it has any memory regardless of the
type of mmemory. If a node has memory then it has at least one zone defined
in its pgdat structure that is located in the pgdat itself.
A node has its bit in N_NORMAL_MEMORY set if it has a lower zone than
ZONE_HIGHMEM. This means it is possible to allocate memory that is not
subject to kmap.
N_HIGH_MEMORY and N_NORMAL_MEMORY can then be used in various places to insure
that we do the right thing when we encounter a memoryless node.
[akpm@linux-foundation.org: build fix]
[Lee.Schermerhorn@hp.com: update N_HIGH_MEMORY node state for memory hotadd]
[y-goto@jp.fujitsu.com: Fix memory hotplug + sparsemem build] Signed-off-by: Lee Schermerhorn <Lee.Schermerhorn@hp.com> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Signed-off-by: Christoph Lameter <clameter@sgi.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com> Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
> For fujitsu, problem is called "empty" node.
>
> When ACPI's SRAT table includes "possible nodes", ia64 bootstrap(acpi_numa_init)
> creates nodes, which includes no memory, no cpu.
>
> I tried to remove empty-node in past, but that was denied.
> It was because we can hot-add cpu to the empty node.
> (node-hotplug triggered by cpu is not implemented now. and it will be ugly.)
>
>
> For HP, (Lee can comment on this later), they have memory-less-node.
> As far as I hear, HP's machine can have following configration.
>
> (example)
> Node0: CPU0 memory AAA MB
> Node1: CPU1 memory AAA MB
> Node2: CPU2 memory AAA MB
> Node3: CPU3 memory AAA MB
> Node4: Memory XXX GB
>
> AAA is very small value (below 16MB) and will be omitted by ia64 bootstrap.
> After boot, only Node 4 has valid memory (but have no cpu.)
>
> Maybe this is memory-interleave by firmware config.
Christoph Lameter <clameter@sgi.com> wrote:
> Future SGI platforms (actually also current one can have but nothing like
> that is deployed to my knowledge) have nodes with only cpus. Current SGI
> platforms have nodes with just I/O that we so far cannot manage in the
> core. So the arch code maps them to the nearest memory node.
Lee Schermerhorn <Lee.Schermerhorn@hp.com> wrote:
> For the HP platforms, we can configure each cell with from 0% to 100%
> "cell local memory". When we configure with <100% CLM, the "missing
> percentages" are interleaved by hardware on a cache-line granularity to
> improve bandwidth at the expense of latency for numa-challenged
> applications [and OSes, but not our problem ;-)]. When we boot Linux on
> such a config, all of the real nodes have no memory--it all resides in a
> single interleaved pseudo-node.
>
> When we boot Linux on a 100% CLM configuration [== NUMA], we still have
> the interleaved pseudo-node. It contains a few hundred MB stolen from
> the real nodes to contain the DMA zone. [Interleaved memory resides at
> phys addr 0]. The memoryless-nodes patches, along with the zoneorder
> patches, support this config as well.
>
> Also, when we boot a NUMA config with the "mem=" command line,
> specifying less memory than actually exists, Linux takes the excluded
> memory "off the top" rather than distributing it across the nodes. This
> can result in memoryless nodes, as well.
>
This patch:
Preparation for memoryless node patches.
Provide a generic way to keep nodemasks describing various characteristics of
NUMA nodes.
Remove the node_online_map and the node_possible map and realize the same
functionality using two nodes stats: N_POSSIBLE and N_ONLINE.
[Lee.Schermerhorn@hp.com: Initialize N_*_MEMORY and N_CPU masks for non-NUMA config] Signed-off-by: Christoph Lameter <clameter@sgi.com> Tested-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Acked-by: Bob Picco <bob.picco@hp.com> Cc: Nishanth Aravamudan <nacc@us.ibm.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@skynet.ie> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:25:26 +0000 (01:25 -0700)]
fs: remove some AOP_TRUNCATED_PAGE
prepare/commit_write no longer returns AOP_TRUNCATED_PAGE since OCFS2 and
GFS2 were converted to the new aops, so we can make some simplifications
for that.
[michal.k.k.piotrowski@gmail.com: fix warning] Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Michael Halcrow <mhalcrow@us.ibm.com> Cc: Mark Fasheh <mark.fasheh@oracle.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Michal Piotrowski <michal.k.k.piotrowski@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:25:25 +0000 (01:25 -0700)]
fs: restore nobh
Implement nobh in new aops. This is a bit tricky. FWIW, nobh_truncate is
now implemented in a way that does not create blocks in sparse regions,
which is a silly thing for it to have been doing (isn't it?)
ext2 survives fsx and fsstress. jfs is converted as well... ext3
should be easy to do (but not done yet).
Nick Piggin [Tue, 16 Oct 2007 08:25:23 +0000 (01:25 -0700)]
fs: adfs convert to new aops
Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> 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>
Nick Piggin [Tue, 16 Oct 2007 08:25:20 +0000 (01:25 -0700)]
udf: convert to new aops
Convert udf to new aops. Also seem to have fixed pagecache corruption in
udf_adinicb_commit_write -- page was marked uptodate when it is not. Also,
fixed the silly setup where prepare_write was doing a kmap to be used in
commit_write: just do kmap_atomic in write_end. Use libfs helpers to make
this easier.
Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: <bfennema@falcon.csc.calpoly.edu> Cc: Jan Kara <jack@ucw.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:25:17 +0000 (01:25 -0700)]
hostfs: convert to new aops
This also gets rid of a lot of useless read_file stuff. And also
optimises the full page write case by marking a !uptodate page uptodate.
Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:25:07 +0000 (01:25 -0700)]
fs: new cont helpers
Rework the generic block "cont" routines to handle the new aops. Supporting
cont_prepare_write would take quite a lot of code to support, so remove it
instead (and we later convert all filesystems to use it).
write_begin gets passed AOP_FLAG_CONT_EXPAND when called from
generic_cont_expand, so filesystems can avoid the old hacks they used.
Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dmitry Monakhov [Tue, 16 Oct 2007 08:25:02 +0000 (01:25 -0700)]
deny partial write for loop dev fd
Partial write can be easily supported by LO_CRYPT_NONE mode, but it is not
easy in LO_CRYPT_CRYPTOAPI case, because of its block nature. I don't know
who still used cryptoapi, but theoretically it is possible. So let's leave
things as they are. Loop device doesn't support partial write before
Nick's "write_begin/write_end" patch set, and let's it behave the same way
after.
Signed-off-by: Dmitriy Monakhov <dmonakhov@openvz.org> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:25:01 +0000 (01:25 -0700)]
fs: introduce write_begin, write_end, and perform_write aops
These are intended to replace prepare_write and commit_write with more
flexible alternatives that are also able to avoid the buffered write
deadlock problems efficiently (which prepare_write is unable to do).
[mark.fasheh@oracle.com: API design contributions, code review and fixes]
[akpm@linux-foundation.org: various fixes]
[dmonakhov@sw.ru: new aop block_write_begin fix] Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com> Signed-off-by: Dmitriy Monakhov <dmonakhov@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:25:00 +0000 (01:25 -0700)]
fs: fix data-loss on error
New buffers against uptodate pages are simply be marked uptodate, while the
buffer_new bit remains set. This causes error-case code to zero out parts of
those buffers because it thinks they contain stale data: wrong, they are
actually uptodate so this is a data loss situation.
Fix this by actually clearning buffer_new and marking the buffer dirty. It
makes sense to always clear buffer_new before setting a buffer uptodate.
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>
Nick Piggin [Tue, 16 Oct 2007 08:24:59 +0000 (01:24 -0700)]
mm: fix pagecache write deadlocks
Modify the core write() code so that it won't take a pagefault while holding a
lock on the pagecache page. There are a number of different deadlocks possible
if we try to do such a thing:
a. sys_munmap / sys_mlock / others
b. mmap_sem(w)
c. make_pages_present
d. get_user_pages
e. handle_mm_fault
f. lock_page (filemap_nopage)
2,8 - recursive deadlock if page is same
2,8;2,8 - ABBA deadlock is page is different
2,6;b,f - ABBA deadlock if page is same
The solution is as follows:
1. If we find the destination page is uptodate, continue as normal, but use
atomic usercopies which do not take pagefaults and do not zero the uncopied
tail of the destination. The destination is already uptodate, so we can
commit_write the full length even if there was a partial copy: it does not
matter that the tail was not modified, because if it is dirtied and written
back to disk it will not cause any problems (uptodate *means* that the
destination page is as new or newer than the copy on disk).
1a. The above requires that fault_in_pages_readable correctly returns access
information, because atomic usercopies cannot distinguish between
non-present pages in a readable mapping, from lack of a readable mapping.
2. If we find the destination page is non uptodate, unlock it (this could be
made slightly more optimal), then allocate a temporary page to copy the
source data into. Relock the destination page and continue with the copy.
However, instead of a usercopy (which might take a fault), copy the data
from the pinned temporary page via the kernel address space.
(also, rename maxlen to seglen, because it was confusing)
This increases the CPU/memory copy cost by almost 50% on the affected
workloads. That will be solved by introducing a new set of pagecache write
aops in a subsequent patch.
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>
Nick Piggin [Tue, 16 Oct 2007 08:24:58 +0000 (01:24 -0700)]
mm: write iovec cleanup
Hide some of the open-coded nr_segs tests into the iovec helpers. This is all
to simplify generic_file_buffered_write, because that gets more complex in the
next patch.
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>
Nick Piggin [Tue, 16 Oct 2007 08:24:57 +0000 (01:24 -0700)]
mm: buffered write cleanup
Quite a bit of code is used in maintaining these "cached pages" that are
probably pretty unlikely to get used. It would require a narrow race where
the page is inserted concurrently while this process is allocating a page
in order to create the spare page. Then a multi-page write into an uncached
part of the file, to make use of it.
Next, the buffered write path (and others) uses its own LRU pagevec when it
should be just using the per-CPU LRU pagevec (which will cut down on both data
and code size cacheline footprint). Also, these private LRU pagevecs are
emptied after just a very short time, in contrast with the per-CPU pagevecs
that are persistent. Net result: 7.3 times fewer lru_lock acquisitions required
to add the pages to pagecache for a bulk write (in 4K chunks).
[this gets rid of some cond_resched() calls in readahead.c and mpage.c due
to clashes in -mm. What put them there, and why? ]
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>
Nick Piggin [Tue, 16 Oct 2007 08:24:56 +0000 (01:24 -0700)]
mm: trim more holes
If prepare_write fails with AOP_TRUNCATED_PAGE, or if commit_write fails, then
we may have failed the write operation despite prepare_write having
instantiated blocks past i_size. Fix this, and consolidate the trimming into
one place.
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>
When prefaulting in the pages in generic_file_buffered_write(), we only
faulted in the pages for the firts segment of the iovec. If the second of
successive segment described a mmapping of the page into which we're
write()ing, and that page is not up-to-date, the fault handler tries to lock
the already-locked page (to bring it up to date) and deadlocks.
An exploit for this bug is in writev-deadlock-demo.c, in
http://www.zip.com.au/~akpm/linux/patches/stuff/ext3-tools.tar.gz.
(These demos assume blocksize < PAGE_CACHE_SIZE).
The problem with this fix is that it takes the kernel back to doing a single
prepare_write()/commit_write() per iovec segment. So in the worst case we'll
run prepare_write+commit_write 1024 times where we previously would have run
it once. The other problem with the fix is that it fix all the locking problems.
<insert numbers obtained via ext3-tools's writev-speed.c here>
And apparently this change killed NFS overwrite performance, because, I
suppose, it talks to the server for each prepare_write+commit_write.
So just back that patch out - we'll be fixing the deadlock by other means.
Nick says: also it only ever actually papered over the bug, because after
faulting in the pages, they might be unmapped or reclaimed.
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>
Nick Piggin [Tue, 16 Oct 2007 08:24:53 +0000 (01:24 -0700)]
mm: revert KERNEL_DS buffered write optimisation
Revert the patch from Neil Brown to optimise NFSD writev handling.
Cc: Neil Brown <neilb@suse.de> Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Hisashi Hifumi [Tue, 16 Oct 2007 08:24:52 +0000 (01:24 -0700)]
mm: use pagevec to rotate reclaimable page
While running some memory intensive load, system response deteriorated just
after swap-out started.
The cause of this problem is that when a PG_reclaim page is moved to the tail
of the inactive LRU list in rotate_reclaimable_page(), lru_lock spin lock is
acquired every page writeback . This deteriorates system performance and
makes interrupt hold off time longer when swap-out started.
Following patch solves this problem. I use pagevec in rotating reclaimable
pages to mitigate LRU spin lock contention and reduce interrupt hold off time.
I did a test that allocating and touching pages in multiple processes, and
pinging to the test machine in flooding mode to measure response under memory
intensive load.
Rik van Riel [Tue, 16 Oct 2007 08:24:50 +0000 (01:24 -0700)]
mm: prevent kswapd from freeing excessive amounts of lowmem
The current VM can get itself into trouble fairly easily on systems with a
small ZONE_HIGHMEM, which is common on i686 computers with 1GB of memory.
On one side, page_alloc() will allocate down to zone->pages_low, while on
the other side, kswapd() and balance_pgdat() will try to free memory from
every zone, until every zone has more free pages than zone->pages_high.
Highmem can be filled up to zone->pages_low with page tables, ramfs,
vmalloc allocations and other unswappable things quite easily and without
many bad side effects, since we still have a huge ZONE_NORMAL to do future
allocations from.
However, as long as the number of free pages in the highmem zone is below
zone->pages_high, kswapd will continue swapping things out from
ZONE_NORMAL, too!
Sami Farin managed to get his system into a stage where kswapd had freed
about 700MB of low memory and was still "going strong".
The attached patch will make kswapd stop paging out data from zones when
there is more than enough memory free. We do go above zone->pages_high in
order to keep pressure between zones equal in normal circumstances, but the
patch should prevent the kind of excesses that made Sami's computer totally
unusable.
Signed-off-by: Rik van Riel <riel@redhat.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
So, if CONFIG_BASE_SMALL is set, you will end up with a
RADIX_TREE_MAX_PATH that is one greater than necessary. The practical
upshot of this is just a bit of wasted memory (one long in the
height_to_maxindex array, an extra pre-allocated radix tree node per
cpu, and extra stack usage in a couple of functions), but it seems
worth getting right.
It's also worth noting that I never build with CONFIG_BASE_SMALL.
What I did to test this was duplicate the code in a small user-space
program and check the results of the calculations for max path and the
contents of the height_to_maxindex array.
Signed-off-by: Jeff Moyer <jmoyer@redhat.com> Acked-by: Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nick Piggin [Tue, 16 Oct 2007 08:24:48 +0000 (01:24 -0700)]
fs: fix nobh error handling
nobh mode error handling is not just pretty slack, it's wrong.
One cannot zero out the whole page to ensure new blocks are zeroed, because
it just brings the whole page "uptodate" with zeroes even if that may not
be the correct uptodate data. Also, other parts of the page may already
contain dirty data which would get lost by zeroing it out. Thirdly, the
writeback of zeroes to the new blocks will also erase existing blocks. All
these conditions are pagecache and/or filesystem corruption.
The problem comes about because we didn't keep track of which buffers
actually are new or old. However it is not enough just to keep only this
state, because at the point we start dirtying parts of the page (new
blocks, with zeroes), the handling of IO errors becomes impossible without
buffers because the page may only be partially uptodate, in which case the
page flags allone cannot capture the state of the parts of the page.
So allocate all buffers for the page upfront, but leave them unattached so
that they don't pick up any other references and can be freed when we're
done. If the error path is hit, then zero the new buffers as the regular
buffer path does, then attach the buffers to the page so that it can
actually be written out correctly and be subject to the normal IO error
handling paths.
As an upshot, we save 1K of kernel stack on ia64 or powerpc 64K page
systems.
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>
Slab allocators: fail if ksize is called with a NULL parameter
A NULL pointer means that the object was not allocated. One cannot
determine the size of an object that has not been allocated. Currently we
return 0 but we really should BUG() on attempts to determine the size of
something nonexistent.
krealloc() interprets NULL to mean a zero sized object. Handle that
separately in krealloc().
Signed-off-by: Christoph Lameter <clameter@sgi.com> Acked-by: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Matt Mackall <mpm@selenic.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dean Nelson [Tue, 16 Oct 2007 08:24:45 +0000 (01:24 -0700)]
calculation of pgoff in do_linear_fault() uses mixed units
The calculation of pgoff in do_linear_fault() should use PAGE_SHIFT and not
PAGE_CACHE_SHIFT since vma->vm_pgoff is in units of PAGE_SIZE and not
PAGE_CACHE_SIZE. At the moment linux/pagemap.h has PAGE_CACHE_SHIFT
defined as PAGE_SHIFT, but should that ever change this calculation would
break.
Signed-off-by: Dean Nelson <dcn@sgi.com> Acked-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Satyam Sharma [Tue, 16 Oct 2007 08:24:44 +0000 (01:24 -0700)]
{slub, slob}: use unlikely() for kfree(ZERO_OR_NULL_PTR) check
Considering kfree(NULL) would normally occur only in error paths and
kfree(ZERO_SIZE_PTR) is uncommon as well, so let's use unlikely() for the
condition check in SLUB's and SLOB's kfree() to optimize for the common
case. SLAB has this already.
Signed-off-by: Satyam Sharma <satyam@infradead.org> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Move the definitions of struct mm_struct and struct vma_area_struct to
include/mm_types.h. This allows to define more function in asm/pgtable.h
and friends with inline assemblies instead of macros. Compile tested on
i386, powerpc, powerpc64, s390-32, s390-64 and x86_64.
[aurelien@aurel32.net: build fix] Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
projects / linux-2.6 / log