Changelog since v2
o Change some function names
o Remove mark_memmap_hole in memmap bring up
o Change CONFIG_SPARSEMEM with CONFIG_ARCH_HAS_HOLES_MEMORYMODEL
I have a plan following as after this patch is acked.
TODO:
1) expand pfn_valid to FALTMEM in ARM
I think we can enhance pfn_valid of FLATMEM in ARM.
Now it is doing binary search and it's expesive.
First of all, After we merge this patch, I expand it to FALTMEM of ARM.
2) remove memmap_valid_within
We can remove memmap_valid_within by strict pfn_valid's tight check.
3) Optimize hole check in sparsemem
In case of spasemem, we can optimize pfn_valid through defining new flag
like SECTION_HAS_HOLE of hole mem_section.
== CUT HERE ==
Kukjin reported oops happen while he change min_free_kbytes
http://www.spinics.net/lists/arm-kernel/msg92894.html
It happen by memory map on sparsemem.
The system has a memory map following as.
section 0 section 1 section 2
0x20000000-0x25000000, 0x40000000-0x50000000, 0x50000000-0x58000000
SECTION_SIZE_BITS 28(256M)
It means section 0 is an incompletely filled section.
Nontheless, current pfn_valid of sparsemem checks pfn loosely.
It checks only mem_section's validation but ARM can free mem_map on hole
to save memory space. So in above case, pfn on 0x25000000 can pass pfn_valid's
validation check. It's not what we want.
We can match section size to smallest valid size.(ex, above case, 16M)
But Russell doesn't like it due to mem_section's memory overhead with different
configuration(ex, 512K section).
I tried to add valid pfn range in mem_section but everyone doesn't like it
due to size overhead. This patch is suggested by KAMEZAWA-san.
I just fixed compile error and change some naming.
This patch registers address of mem_section to memmap itself's page struct's
pg->private field. This means the page is used for memmap of the section.
Otherwise, the page is used for other purpose and memmap has a hole.
This patch is based on mmotm-2010-07-19
Signed-off-by: Minchan Kim <[email protected]>
Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
Reported-by: Kukjin Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Russell King <[email protected]>
---
arch/arm/mm/init.c | 4 +++-
include/linux/mmzone.h | 22 +++++++++++++++++++++-
mm/mmzone.c | 34 ++++++++++++++++++++++++++++++++++
5 files changed, 58 insertions(+), 2 deletions(-)
diff --git a/arch/arm/mm/init.c b/arch/arm/mm/init.c
index f6a9994..25e2670 100644
--- a/arch/arm/mm/init.c
+++ b/arch/arm/mm/init.c
@@ -482,8 +482,10 @@ free_memmap(int node, unsigned long start_pfn, unsigned long end_pfn)
* If there are free pages between these,
* free the section of the memmap array.
*/
- if (pg < pgend)
+ if (pg < pgend) {
+ mark_invalid_memmap(pg >> PAGE_SHIFT, pgend >> PAGE_SHIFT);
free_bootmem_node(NODE_DATA(node), pg, pgend - pg);
+ }
}
/*
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index b4d109e..a3195bd 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -15,6 +15,7 @@
#include <linux/seqlock.h>
#include <linux/nodemask.h>
#include <linux/pageblock-flags.h>
+#include <linux/mm_types.h>
#include <generated/bounds.h>
#include <asm/atomic.h>
#include <asm/page.h>
@@ -1049,11 +1050,30 @@ static inline struct mem_section *__pfn_to_section(unsigned long pfn)
return __nr_to_section(pfn_to_section_nr(pfn));
}
+void mark_invalid_memmap(unsigned long start, unsigned long end);
+
+#ifdef CONFIG_ARCH_HAS_HOLES_MEMORYMODEL
+#define MEMMAP_HOLE (0x1UL)
+static inline int memmap_valid(unsigned long pfn)
+{
+ struct page *page = pfn_to_page(pfn);
+ struct page *__pg = virt_to_page(page);
+ return !(__pg->private & MEMMAP_HOLE);
+}
+#else
+static inline int memmap_valid(unsigned long pfn)
+{
+ return 1;
+}
+#endif
+
static inline int pfn_valid(unsigned long pfn)
{
+ struct mem_section *ms;
if (pfn_to_section_nr(pfn) >= NR_MEM_SECTIONS)
return 0;
- return valid_section(__nr_to_section(pfn_to_section_nr(pfn)));
+ ms = __nr_to_section(pfn_to_section_nr(pfn));
+ return valid_section(ms) && memmap_valid(pfn);
}
static inline int pfn_present(unsigned long pfn)
diff --git a/mm/mmzone.c b/mm/mmzone.c
index f5b7d17..7c84e5e 100644
--- a/mm/mmzone.c
+++ b/mm/mmzone.c
@@ -86,4 +86,38 @@ int memmap_valid_within(unsigned long pfn,
return 1;
}
+
+/*
+ * Fill pg->private on hole memmap with MEMMAP_HOLE.
+ * pfn_valid() will check this later. (see include/linux/mmzone.h)
+ * Evenry arch should call
+ * mark_invalid_memmap(start, end) # for all holes in mem_map.
+ * please see usage in ARM.
+ */
+void mark_invalid_memmap(unsigned long start, unsigned long end)
+{
+ struct mem_section *ms;
+ unsigned long pos, next;
+ struct page *pg;
+ void *memmap, *mapend;
+
+ for (pos = start; pos < end; pos = next) {
+ next = (pos + PAGES_PER_SECTION) & PAGE_SECTION_MASK;
+ ms = __pfn_to_section(pos);
+ if (!valid_section(ms))
+ continue;
+
+ for (memmap = (void*)pfn_to_page(pos),
+ /* The last page in section */
+ mapend = pfn_to_page(next-1);
+ memmap < mapend; memmap += PAGE_SIZE) {
+ pg = virt_to_page(memmap);
+ pg->private = MEMMAP_HOLE;
+ }
+ }
+}
+#else
+void mark_invalid_memmap(unsigned long start, unsigned long end)
+{
+}
#endif /* CONFIG_ARCH_HAS_HOLES_MEMORYMODEL */
--
1.7.0.5
On Sun, Jul 25, 2010 at 7:32 PM, Minchan Kim <[email protected]> wrote:
> Changelog since v2
> ?o Change some function names
> ?o Remove mark_memmap_hole in memmap bring up
> ?o Change CONFIG_SPARSEMEM with CONFIG_ARCH_HAS_HOLES_MEMORYMODEL
>
> I have a plan following as after this patch is acked.
>
> TODO:
> 1) expand pfn_valid to FALTMEM in ARM
> I think we can enhance pfn_valid of FLATMEM in ARM.
> Now it is doing binary search and it's expesive.
> First of all, After we merge this patch, I expand it to FALTMEM of ARM.
>
> 2) remove memmap_valid_within
> We can remove memmap_valid_within by strict pfn_valid's tight check.
>
> 3) Optimize hole check in sparsemem
> In case of spasemem, we can optimize pfn_valid through defining new flag
> like SECTION_HAS_HOLE of hole mem_section.
>
Is there an assumption somewhere that assumes that page->private will
always have MEMMAP_HOLE set when the pfn is invalid, independent of
the context in which it is invoked? BTW, I'd also recommend moving
over to using set_page_private() and page_private() wrappers (makes
the code easier to search)
Balbir
> Changelog since v2
> o Change some function names
> o Remove mark_memmap_hole in memmap bring up
> o Change CONFIG_SPARSEMEM with CONFIG_ARCH_HAS_HOLES_MEMORYMODEL
>
> I have a plan following as after this patch is acked.
>
> TODO:
> 1) expand pfn_valid to FALTMEM in ARM
> I think we can enhance pfn_valid of FLATMEM in ARM.
> Now it is doing binary search and it's expesive.
> First of all, After we merge this patch, I expand it to FALTMEM of ARM.
>
> 2) remove memmap_valid_within
> We can remove memmap_valid_within by strict pfn_valid's tight check.
>
> 3) Optimize hole check in sparsemem
> In case of spasemem, we can optimize pfn_valid through defining new
> flag
> like SECTION_HAS_HOLE of hole mem_section.
>
> == CUT HERE ==
>
> Kukjin reported oops happen while he change min_free_kbytes
> http://www.spinics.net/lists/arm-kernel/msg92894.html
> It happen by memory map on sparsemem.
>
> The system has a memory map following as.
> section 0 section 1 section 2
> 0x20000000-0x25000000, 0x40000000-0x50000000, 0x50000000-0x58000000
> SECTION_SIZE_BITS 28(256M)
>
> It means section 0 is an incompletely filled section.
> Nontheless, current pfn_valid of sparsemem checks pfn loosely.
> It checks only mem_section's validation but ARM can free mem_map on
> hole
> to save memory space. So in above case, pfn on 0x25000000 can pass
> pfn_valid's
> validation check. It's not what we want.
>
> We can match section size to smallest valid size.(ex, above case, 16M)
> But Russell doesn't like it due to mem_section's memory overhead with
> different
> configuration(ex, 512K section).
>
> I tried to add valid pfn range in mem_section but everyone doesn't like
> it
> due to size overhead. This patch is suggested by KAMEZAWA-san.
> I just fixed compile error and change some naming.
>
> This patch registers address of mem_section to memmap itself's page
> struct's
> pg->private field. This means the page is used for memmap of the
> section.
> Otherwise, the page is used for other purpose and memmap has a hole.
>
> This patch is based on mmotm-2010-07-19
>
> Signed-off-by: Minchan Kim <[email protected]>
> Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
> Reported-by: Kukjin Kim <[email protected]>
> Cc: Mel Gorman <[email protected]>
> Cc: Johannes Weiner <[email protected]>
> Cc: Russell King <[email protected]>
> ---
> arch/arm/mm/init.c | 4 +++-
> include/linux/mmzone.h | 22 +++++++++++++++++++++-
> mm/mmzone.c | 34
> ++++++++++++++++++++++++++++++++++
> 5 files changed, 58 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm/mm/init.c b/arch/arm/mm/init.c
> index f6a9994..25e2670 100644
> --- a/arch/arm/mm/init.c
> +++ b/arch/arm/mm/init.c
> @@ -482,8 +482,10 @@ free_memmap(int node, unsigned long start_pfn,
> unsigned long end_pfn)
> * If there are free pages between these,
> * free the section of the memmap array.
> */
> - if (pg < pgend)
> + if (pg < pgend) {
> + mark_invalid_memmap(pg >> PAGE_SHIFT, pgend >> PAGE_SHIFT);
> free_bootmem_node(NODE_DATA(node), pg, pgend - pg);
> + }
> }
>
> /*
> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
> index b4d109e..a3195bd 100644
> --- a/include/linux/mmzone.h
> +++ b/include/linux/mmzone.h
> @@ -15,6 +15,7 @@
> #include <linux/seqlock.h>
> #include <linux/nodemask.h>
> #include <linux/pageblock-flags.h>
> +#include <linux/mm_types.h>
> #include <generated/bounds.h>
> #include <asm/atomic.h>
> #include <asm/page.h>
> @@ -1049,11 +1050,30 @@ static inline struct mem_section
> *__pfn_to_section(unsigned long pfn)
> return __nr_to_section(pfn_to_section_nr(pfn));
> }
>
> +void mark_invalid_memmap(unsigned long start, unsigned long end);
> +
> +#ifdef CONFIG_ARCH_HAS_HOLES_MEMORYMODEL
> +#define MEMMAP_HOLE (0x1UL)
> +static inline int memmap_valid(unsigned long pfn)
> +{
> + struct page *page = pfn_to_page(pfn);
> + struct page *__pg = virt_to_page(page);
> + return !(__pg->private & MEMMAP_HOLE);
> +}
> +#else
> +static inline int memmap_valid(unsigned long pfn)
> +{
> + return 1;
> +}
> +#endif
> +
> static inline int pfn_valid(unsigned long pfn)
> {
> + struct mem_section *ms;
> if (pfn_to_section_nr(pfn) >= NR_MEM_SECTIONS)
> return 0;
> - return valid_section(__nr_to_section(pfn_to_section_nr(pfn)));
> + ms = __nr_to_section(pfn_to_section_nr(pfn));
> + return valid_section(ms) && memmap_valid(pfn);
> }
>
> static inline int pfn_present(unsigned long pfn)
> diff --git a/mm/mmzone.c b/mm/mmzone.c
> index f5b7d17..7c84e5e 100644
> --- a/mm/mmzone.c
> +++ b/mm/mmzone.c
> @@ -86,4 +86,38 @@ int memmap_valid_within(unsigned long pfn,
>
> return 1;
> }
> +
> +/*
> + * Fill pg->private on hole memmap with MEMMAP_HOLE.
> + * pfn_valid() will check this later. (see include/linux/mmzone.h)
> + * Evenry arch should call
> + * mark_invalid_memmap(start, end) # for all holes in mem_map.
> + * please see usage in ARM.
> + */
> +void mark_invalid_memmap(unsigned long start, unsigned long end)
> +{
> + struct mem_section *ms;
> + unsigned long pos, next;
> + struct page *pg;
> + void *memmap, *mapend;
> +
> + for (pos = start; pos < end; pos = next) {
> + next = (pos + PAGES_PER_SECTION) & PAGE_SECTION_MASK;
> + ms = __pfn_to_section(pos);
> + if (!valid_section(ms))
> + continue;
> +
> + for (memmap = (void*)pfn_to_page(pos),
> + /* The last page in section */
> + mapend = pfn_to_page(next-1);
> + memmap < mapend; memmap += PAGE_SIZE) {
> + pg = virt_to_page(memmap);
> + pg->private = MEMMAP_HOLE;
> + }
> + }
> +}
> +#else
> +void mark_invalid_memmap(unsigned long start, unsigned long end)
> +{
> +}
> #endif /* CONFIG_ARCH_HAS_HOLES_MEMORYMODEL */
> --
I don't think this works because if the memmap pages get reused, the corresponding struct page->private could be used by chance in such a way that it has the value of MEMMAP_HOLE. Of course unlikely but possible. And after all the whole point of freeing part of the memmap is that it could be reused.
--Mika
On Mon, Jul 26, 2010 at 2:03 AM, Balbir Singh <[email protected]> wrote:
> On Sun, Jul 25, 2010 at 7:32 PM, Minchan Kim <[email protected]> wrote:
>> Changelog since v2
>> ?o Change some function names
>> ?o Remove mark_memmap_hole in memmap bring up
>> ?o Change CONFIG_SPARSEMEM with CONFIG_ARCH_HAS_HOLES_MEMORYMODEL
>>
>> I have a plan following as after this patch is acked.
>>
>> TODO:
>> 1) expand pfn_valid to FALTMEM in ARM
>> I think we can enhance pfn_valid of FLATMEM in ARM.
>> Now it is doing binary search and it's expesive.
>> First of all, After we merge this patch, I expand it to FALTMEM of ARM.
>>
>> 2) remove memmap_valid_within
>> We can remove memmap_valid_within by strict pfn_valid's tight check.
>>
>> 3) Optimize hole check in sparsemem
>> In case of spasemem, we can optimize pfn_valid through defining new flag
>> like SECTION_HAS_HOLE of hole mem_section.
>>
>
> Is there an assumption somewhere that assumes that page->private will
> always have MEMMAP_HOLE set when the pfn is invalid, independent of
> the context in which it is invoked? BTW, I'd also recommend moving
zzzzz.... I needed sleep.
Will fix and resend.
> over to using set_page_private() and page_private() wrappers (makes
> the code easier to search)
Okay.
Thanks for pointing me out, Balbir.
>
> Balbir
>
--
Kind regards,
Minchan Kim
2010/7/26 Penttil? Mika <[email protected]>:
> I don't think this works because if the memmap pages get reused, the corresponding struct page->private could be used by chance in such a way that it has the value of MEMMAP_HOLE. Of course unlikely but possible. And after all the whole point of freeing part of the memmap is that it could be reused.
>
You're absolutely right.
Previous version, I didn't do such as.
In this version, I wanted to remove dependency of page->private and
mem_section to identify hole memmap for using it in FLATMEM of ARM but
make mistake.
I will resend the patch
Thanks for careful review.
> --Mika
>
>
>
>
--
Kind regards,
Minchan Kim