mm: setup pageblock_order before it's used by sparsemem
authorXishi Qiu <qiuxishi@huawei.com>
Tue, 31 Jul 2012 23:43:19 +0000 (16:43 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 1 Aug 2012 01:42:43 +0000 (18:42 -0700)
On architectures with CONFIG_HUGETLB_PAGE_SIZE_VARIABLE set, such as
Itanium, pageblock_order is a variable with default value of 0.  It's set
to the right value by set_pageblock_order() in function
free_area_init_core().

But pageblock_order may be used by sparse_init() before free_area_init_core()
is called along path:
sparse_init()
    ->sparse_early_usemaps_alloc_node()
->usemap_size()
    ->SECTION_BLOCKFLAGS_BITS
->((1UL << (PFN_SECTION_SHIFT - pageblock_order)) *
NR_PAGEBLOCK_BITS)

The uninitialized pageblock_size will cause memory wasting because
usemap_size() returns a much bigger value then it's really needed.

For example, on an Itanium platform,
sparse_init() pageblock_order=0 usemap_size=24576
free_area_init_core() before pageblock_order=0, usemap_size=24576
free_area_init_core() after pageblock_order=12, usemap_size=8

That means 24K memory has been wasted for each section, so fix it by calling
set_pageblock_order() from sparse_init().

Signed-off-by: Xishi Qiu <qiuxishi@huawei.com>
Signed-off-by: Jiang Liu <liuj97@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Keping Chen <chenkeping@huawei.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/internal.h
mm/page_alloc.c
mm/sparse.c

index da6b9b2ed3fc429443add0dc2349098b83ee9ea8..3314f79d775a5f90c84e3250d8b93b7bd6f99933 100644 (file)
@@ -353,3 +353,5 @@ extern u32 hwpoison_filter_enable;
 extern unsigned long vm_mmap_pgoff(struct file *, unsigned long,
         unsigned long, unsigned long,
         unsigned long, unsigned long);
+
+extern void set_pageblock_order(void);
index 94fc475c3f94ec72e116e2e2b327a5b1ecd1330d..6c7e3bd93a856d9d1d46ae4ca6062f298be6e43f 100644 (file)
@@ -4304,7 +4304,7 @@ static inline void setup_usemap(struct pglist_data *pgdat,
 #ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
 
 /* Initialise the number of pages represented by NR_PAGEBLOCK_BITS */
-static inline void __init set_pageblock_order(void)
+void __init set_pageblock_order(void)
 {
        unsigned int order;
 
@@ -4332,7 +4332,7 @@ static inline void __init set_pageblock_order(void)
  * include/linux/pageblock-flags.h for the values of pageblock_order based on
  * the kernel config
  */
-static inline void set_pageblock_order(void)
+void __init set_pageblock_order(void)
 {
 }
 
index c7bb952400c83c9114a401f647955cb18dd2ea44..950981fd07c5f2686b08cb864c2e0f3d548d4d95 100644 (file)
@@ -493,6 +493,9 @@ void __init sparse_init(void)
        struct page **map_map;
 #endif
 
+       /* Setup pageblock_order for HUGETLB_PAGE_SIZE_VARIABLE */
+       set_pageblock_order();
+
        /*
         * map is using big page (aka 2M in x86 64 bit)
         * usemap is less one page (aka 24 bytes)