linux-stable.git/mm/memblock.c, branch linux-3.4.y

x86, mm: Trim memory in memblock to be page aligned

2012-10-31T17:02:56+00:00

commit 6ede1fd3cb404c0016de6ac529df46d561bd558b upstream.

We will not map partial pages, so need to make sure memblock
allocation will not allocate those bytes out.

Also we will use for_each_mem_pfn_range() to loop to map memory
range to keep them consistent.

Signed-off-by: Yinghai Lu 
Link: http://lkml.kernel.org/r/CAE9FiQVZirvaBMFYRfXMmWEcHbKSicQEHz4VAwUv0xFCk51ZNw@mail.gmail.com
Acked-by: Jacob Shin 
Signed-off-by: H. Peter Anvin 
Signed-off-by: Greg Kroah-Hartman

memblock: free allocated memblock_reserved_regions later

2012-07-16T16:04:45+00:00

commit 29f6738609e40227dabcc63bfb3b84b3726a75bd upstream.

memblock_free_reserved_regions() calls memblock_free(), but
memblock_free() would double reserved.regions too, so we could free the
old range for reserved.regions.

Also tj said there is another bug which could be related to this.

| I don't think we're saving any noticeable
| amount by doing this "free - give it to page allocator - reserve
| again" dancing.  We should just allocate regions aligned to page
| boundaries and free them later when memblock is no longer in use.

in that case, when DEBUG_PAGEALLOC, will get panic:

     memblock_free: [0x0000102febc080-0x0000102febf080] memblock_free_reserved_regions+0x37/0x39
  BUG: unable to handle kernel paging request at ffff88102febd948
  IP: [] __next_free_mem_range+0x9b/0x155
  PGD 4826063 PUD cf67a067 PMD cf7fa067 PTE 800000102febd160
  Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
  CPU 0
  Pid: 0, comm: swapper Not tainted 3.5.0-rc2-next-20120614-sasha #447
  RIP: 0010:[]  [] __next_free_mem_range+0x9b/0x155

See the discussion at https://lkml.org/lkml/2012/6/13/469

So try to allocate with PAGE_SIZE alignment and free it later.

Reported-by: Sasha Levin 
Acked-by: Tejun Heo 
Cc: Benjamin Herrenschmidt 
Signed-off-by: Yinghai Lu 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

mm/memblock: fix overlapping allocation when doubling reserved array

2012-07-16T16:04:27+00:00

commit 48c3b583bbddad2220ca4c22319ca5d1f78b2090 upstream.

__alloc_memory_core_early() asks memblock for a range of memory then try
to reserve it.  If the reserved region array lacks space for the new
range, memblock_double_array() is called to allocate more space for the
array.  If memblock is used to allocate memory for the new array it can
end up using a range that overlaps with the range originally allocated in
__alloc_memory_core_early(), leading to possible data corruption.

With this patch memblock_double_array() now calls memblock_find_in_range()
with a narrowed candidate range (in cases where the reserved.regions array
is being doubled) so any memory allocated will not overlap with the
original range that was being reserved.  The range is narrowed by passing
in the starting address and size of the previously allocated range.  Then
the range above the ending address is searched and if a candidate is not
found, the range below the starting address is searched.

Signed-off-by: Greg Pearson 
Signed-off-by: Yinghai Lu 
Acked-by: Tejun Heo 
Cc: Benjamin Herrenschmidt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

mm/memblock: fix memory leak on extending regions

2012-07-16T16:04:27+00:00

commit 181eb39425f2b9275afcb015eaa547d11f71a02f upstream.

The overall memblock has been organized into the memory regions and
reserved regions.  Initially, the memory regions and reserved regions are
stored in the predetermined arrays of "struct memblock _region".  It's
possible for the arrays to be enlarged when we have newly added regions,
but no free space left there.  The policy here is to create double-sized
array either by slab allocator or memblock allocator.  Unfortunately, we
didn't free the old array, which might be allocated through slab allocator
before.  That would cause memory leak.

The patch introduces 2 variables to trace where (slab or memblock) the
memory and reserved regions come from.  The memory for the memory or
reserved regions will be deallocated by kfree() if that was allocated by
slab allocator.  Thus to fix the memory leak issue.

Signed-off-by: Gavin Shan 
Cc: Johannes Weiner 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

mm/memblock: cleanup on duplicate VA/PA conversion

2012-07-16T16:04:27+00:00

commit 4e2f07750d9a94e8f23e86408df5ab95be88bf11 upstream.

The overall memblock has been organized into the memory regions and
reserved regions.  Initially, the memory regions and reserved regions are
stored in the predetermined arrays of "struct memblock _region".  It's
possible for the arrays to be enlarged when we have newly added regions
for them, but no enough space there.  Under the situation, We will created
double-sized array to meet the requirement.  However, the original
implementation converted the VA (Virtual Address) of the newly allocated
array of regions to PA (Physical Address), then translate back when we
allocates the new array from slab.  That's actually unnecessary.

The patch removes the duplicate VA/PA conversion.

Signed-off-by: Gavin Shan 
Cc: Johannes Weiner 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

memblock: memblock should be able to handle zero length operations

2012-04-20T18:18:46+00:00

Commit 24aa07882b ("memblock, x86: Replace memblock_x86_reserve/
free_range() with generic ones") replaced x86 specific memblock
operations with the generic ones; unfortunately, it lost zero length
operation handling in the process making the kernel panic if somebody
tries to reserve zero length area.

There isn't much to be gained by being cranky to zero length operations
and panicking is almost the worst response.  Drop the BUG_ON() in
memblock_reserve() and update memblock_add_region/isolate_range() so
that all zero length operations are handled as noops.

Signed-off-by: Tejun Heo 
Cc: stable@vger.kernel.org
Reported-by: Valere Monseur 
Bisected-by: Joseph Freeman 
Tested-by: Joseph Freeman 
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=43098
Signed-off-by: Linus Torvalds

memblock: Fix size aligning of memblock_alloc_base_nid()

2012-03-01T09:53:18+00:00

memblock allocator aligns @size to @align to reduce the amount
of fragmentation.  Commit:

 7bd0b0f0da ("memblock: Reimplement memblock allocation using reverse free area iterator")

Broke it by incorrectly relocating @size aligning to
memblock_find_in_range_node().  As the aligned size is not
propagated back to memblock_alloc_base_nid(), the actually
reserved size isn't aligned.

While this increases memory use for memblock reserved array,
this shouldn't cause any critical failure; however, it seems
that the size aligning was hiding a use-beyond-allocation bug in
sparc64 and losing the aligning causes boot failure.

The underlying problem is currently being debugged but this is a
proper fix in itself, it's already pretty late in -rc cycle for
boot failures and reverting the change for debugging isn't
difficult. Restore the size aligning moving it to
memblock_alloc_base_nid().

Reported-by: Meelis Roos 
Signed-off-by: Tejun Heo 
Cc: David S. Miller 
Cc: Grant Likely 
Cc: Rob Herring 
Cc: Linus Torvalds 
Cc: Andrew Morton 
Link: http://lkml.kernel.org/r/20120228205621.GC3252@dhcp-172-17-108-109.mtv.corp.google.com
Signed-off-by: Ingo Molnar 
LKML-Reference:

memblock: Fix alloc failure due to dumb underflow protection in memblock_find_in_range_node()

2012-01-16T07:38:06+00:00

7bd0b0f0da ("memblock: Reimplement memblock allocation using
reverse free area iterator") implemented a simple top-down
allocator using a reverse memblock iterator.  To avoid underflow
in the allocator loop, it simply raised the lower boundary to
the requested size under the assumption that requested size
would be far smaller than available memblocks.

This causes early page table allocation failure under certain
configurations in Xen.  Fix it by checking for underflow directly
instead of bumping up lower bound.

Signed-off-by: Tejun Heo 
Reported-by: Konrad Rzeszutek Wilk 
Cc: rjw@sisk.pl
Cc: xen-devel@lists.xensource.com
Cc: Benjamin Herrenschmidt 
Cc: Andrew Morton 
Cc: Linus Torvalds 
Link: http://lkml.kernel.org/r/20120113181412.GA11112@google.com
Signed-off-by: Ingo Molnar

memblock: Reimplement memblock allocation using reverse free area iterator

2011-12-08T18:22:09+00:00

Now that all early memory information is in memblock when enabled, we
can implement reverse free area iterator and use it to implement NUMA
aware allocator which is then wrapped for simpler variants instead of
the confusing and inefficient mending of information in separate NUMA
aware allocator.

Implement for_each_free_mem_range_reverse(), use it to reimplement
memblock_find_in_range_node() which in turn is used by all allocators.

The visible allocator interface is inconsistent and can probably use
some cleanup too.

Signed-off-by: Tejun Heo 
Cc: Benjamin Herrenschmidt 
Cc: Yinghai Lu

memblock: Kill early_node_map[]

2011-12-08T18:22:09+00:00

Now all ARCH_POPULATES_NODE_MAP archs select HAVE_MEBLOCK_NODE_MAP -
there's no user of early_node_map[] left.  Kill early_node_map[] and
replace ARCH_POPULATES_NODE_MAP with HAVE_MEMBLOCK_NODE_MAP.  Also,
relocate for_each_mem_pfn_range() and helper from mm.h to memblock.h
as page_alloc.c would no longer host an alternative implementation.

This change is ultimately one to one mapping and shouldn't cause any
observable difference; however, after the recent changes, there are
some functions which now would fit memblock.c better than page_alloc.c
and dependency on HAVE_MEMBLOCK_NODE_MAP instead of HAVE_MEMBLOCK
doesn't make much sense on some of them.  Further cleanups for
functions inside HAVE_MEMBLOCK_NODE_MAP in mm.h would be nice.

-v2: Fix compile bug introduced by mis-spelling
 CONFIG_HAVE_MEMBLOCK_NODE_MAP to CONFIG_MEMBLOCK_HAVE_NODE_MAP in
 mmzone.h.  Reported by Stephen Rothwell.

Signed-off-by: Tejun Heo 
Cc: Stephen Rothwell 
Cc: Benjamin Herrenschmidt 
Cc: Yinghai Lu 
Cc: Tony Luck 
Cc: Ralf Baechle 
Cc: Martin Schwidefsky 
Cc: Chen Liqin 
Cc: Paul Mundt 
Cc: "David S. Miller" 
Cc: "H. Peter Anvin"