mm: use aligned zone start for pfn_to_bitidx calculation
authorLaura Abbott <lauraa@codeaurora.org>
Fri, 11 Jan 2013 22:31:51 +0000 (14:31 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 11 Jan 2013 22:54:55 +0000 (14:54 -0800)
commitc060f943d0929f3e429c5d9522290584f6281d6e
treeccbc72cee7202d8591aa9c5fa55dcc465bbd2a35
parent6d92d4f6a74766cc885b18218268e0c47fbca399
mm: use aligned zone start for pfn_to_bitidx calculation

The current calculation in pfn_to_bitidx assumes that (pfn -
zone->zone_start_pfn) >> pageblock_order will return the same bit for
all pfn in a pageblock.  If zone_start_pfn is not aligned to
pageblock_nr_pages, this may not always be correct.

Consider the following with pageblock order = 10, zone start 2MB:

  pfn     | pfn - zone start | (pfn - zone start) >> page block order
  ----------------------------------------------------------------
  0x26000 | 0x25e00    |  0x97
  0x26100 | 0x25f00    |  0x97
  0x26200 | 0x26000    |  0x98
  0x26300 | 0x26100    |  0x98

This means that calling {get,set}_pageblock_migratetype on a single page
will not set the migratetype for the full block.  Fix this by rounding
down zone_start_pfn when doing the bitidx calculation.

For our use case, the effects of this bug were mostly tied to the fact
that CMA allocations would either take a long time or fail to happen.
Depending on the driver using CMA, this could result in anything from
visual glitches to application failures.

Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
Acked-by: Mel Gorman <mgorman@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/page_alloc.c