From: Xishi Qiu <[email protected]>
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 <[email protected]>
Signed-off-by: Jiang Liu <[email protected]>
---
mm/internal.h | 2 ++
mm/page_alloc.c | 4 ++--
mm/sparse.c | 3 +++
3 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/mm/internal.h b/mm/internal.h
index 2ba87fb..8052379 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -347,3 +347,5 @@ 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);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4403009..f38509b 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4301,7 +4301,7 @@ 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;
@@ -4329,7 +4329,7 @@ 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)
{
}
diff --git a/mm/sparse.c b/mm/sparse.c
index fca2ab5..3a3af73 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -485,6 +485,9 @@ 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)
--
1.7.1
On Sat, Jun 30, 2012 at 2:07 AM, Jiang Liu <[email protected]> wrote:
> From: Xishi Qiu <[email protected]>
>
> 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().
>
can you check attached patch?
That will kill more lines code instead.
Thanks
Yinghai
On Sat, Jun 30, 2012 at 05:07:54PM +0800, Jiang Liu wrote:
> From: Xishi Qiu <[email protected]>
>
> 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 <[email protected]>
> Signed-off-by: Jiang Liu <[email protected]>
> ---
> mm/internal.h | 2 ++
> mm/page_alloc.c | 4 ++--
> mm/sparse.c | 3 +++
> 3 files changed, 7 insertions(+), 2 deletions(-)
<formletter>
This is not the correct way to submit patches for inclusion in the
stable kernel tree. Please read Documentation/stable_kernel_rules.txt
for how to do this properly.
</formletter>