All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init()
@ 2018-11-22  9:48 Wei Yang
  2018-12-20 23:01 ` Andrew Morton
  2018-12-20 23:50 ` Alexander Duyck
  0 siblings, 2 replies; 4+ messages in thread
From: Wei Yang @ 2018-11-22  9:48 UTC (permalink / raw)
  To: osalvador; +Cc: akpm, linux-mm, Wei Yang

When DEFERRED_STRUCT_PAGE_INIT is configured, only the first section of
each node's highest zone is initialized before defer stage.

static_init_pgcnt is used to store the number of pages like this:

    pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
                                              pgdat->node_spanned_pages);

because we don't want to overflow zone's range.

But this is not necessary, since defer_init() is called like this:

  memmap_init_zone()
    for pfn in [start_pfn, end_pfn)
      defer_init(pfn, end_pfn)

In case (pgdat->node_spanned_pages < PAGES_PER_SECTION), the loop would
stop before calling defer_init().

BTW, comparing PAGES_PER_SECTION with node_spanned_pages is not correct,
since nr_initialised is zone based instead of node based. Even
node_spanned_pages is bigger than PAGES_PER_SECTION, its highest zone
would have pages less than PAGES_PER_SECTION.

Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
---
 include/linux/mmzone.h |  2 --
 mm/page_alloc.c        | 13 ++++++-------
 2 files changed, 6 insertions(+), 9 deletions(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index ae7a830a21eb..68d7b558924b 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -692,8 +692,6 @@ typedef struct pglist_data {
 	 * is the first PFN that needs to be initialised.
 	 */
 	unsigned long first_deferred_pfn;
-	/* Number of non-deferred pages */
-	unsigned long static_init_pgcnt;
 #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 76e12179cd5e..b542d82400cf 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -326,8 +326,13 @@ defer_init(int nid, unsigned long pfn, unsigned long end_pfn)
 	/* Always populate low zones for address-constrained allocations */
 	if (end_pfn < pgdat_end_pfn(NODE_DATA(nid)))
 		return false;
+
+	/*
+	 * We start only with one section of pages, more pages are added as
+	 * needed until the rest of deferred pages are initialized.
+	 */
 	nr_initialised++;
-	if ((nr_initialised > NODE_DATA(nid)->static_init_pgcnt) &&
+	if ((nr_initialised > PAGES_PER_SECTION) &&
 	    (pfn & (PAGES_PER_SECTION - 1)) == 0) {
 		NODE_DATA(nid)->first_deferred_pfn = pfn;
 		return true;
@@ -6451,12 +6456,6 @@ static void __ref alloc_node_mem_map(struct pglist_data *pgdat) { }
 #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
 static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
 {
-	/*
-	 * We start only with one section of pages, more pages are added as
-	 * needed until the rest of deferred pages are initialized.
-	 */
-	pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
-						pgdat->node_spanned_pages);
 	pgdat->first_deferred_pfn = ULONG_MAX;
 }
 #else
-- 
2.15.1

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init()
  2018-11-22  9:48 [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init() Wei Yang
@ 2018-12-20 23:01 ` Andrew Morton
  2018-12-20 23:50 ` Alexander Duyck
  1 sibling, 0 replies; 4+ messages in thread
From: Andrew Morton @ 2018-12-20 23:01 UTC (permalink / raw)
  To: Wei Yang; +Cc: osalvador, linux-mm


Could someone please review this?



From: Wei Yang <richard.weiyang@gmail.com>
Subject: mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init()

When DEFERRED_STRUCT_PAGE_INIT is configured, only the first section of
each node's highest zone is initialized before defer stage.

static_init_pgcnt is used to store the number of pages like this:

    pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
                                              pgdat->node_spanned_pages);

because we don't want to overflow zone's range.

But this is not necessary, since defer_init() is called like this:

  memmap_init_zone()
    for pfn in [start_pfn, end_pfn)
      defer_init(pfn, end_pfn)

In case (pgdat->node_spanned_pages < PAGES_PER_SECTION), the loop would
stop before calling defer_init().

BTW, comparing PAGES_PER_SECTION with node_spanned_pages is not correct,
since nr_initialised is zone based instead of node based.  Even
node_spanned_pages is bigger than PAGES_PER_SECTION, its highest zone
would have pages less than PAGES_PER_SECTION.

Link: http://lkml.kernel.org/r/20181122094807.6985-1-richard.weiyang@gmail.com
Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 include/linux/mmzone.h |    2 --
 mm/page_alloc.c        |   13 ++++++-------
 2 files changed, 6 insertions(+), 9 deletions(-)

--- a/include/linux/mmzone.h~mm-check-nr_initialised-with-pages_per_section-directly-in-defer_init
+++ a/include/linux/mmzone.h
@@ -692,8 +692,6 @@ typedef struct pglist_data {
 	 * is the first PFN that needs to be initialised.
 	 */
 	unsigned long first_deferred_pfn;
-	/* Number of non-deferred pages */
-	unsigned long static_init_pgcnt;
 #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
--- a/mm/page_alloc.c~mm-check-nr_initialised-with-pages_per_section-directly-in-defer_init
+++ a/mm/page_alloc.c
@@ -326,8 +326,13 @@ defer_init(int nid, unsigned long pfn, u
 	/* Always populate low zones for address-constrained allocations */
 	if (end_pfn < pgdat_end_pfn(NODE_DATA(nid)))
 		return false;
+
+	/*
+	 * We start only with one section of pages, more pages are added as
+	 * needed until the rest of deferred pages are initialized.
+	 */
 	nr_initialised++;
-	if ((nr_initialised > NODE_DATA(nid)->static_init_pgcnt) &&
+	if ((nr_initialised > PAGES_PER_SECTION) &&
 	    (pfn & (PAGES_PER_SECTION - 1)) == 0) {
 		NODE_DATA(nid)->first_deferred_pfn = pfn;
 		return true;
@@ -6585,12 +6590,6 @@ static void __ref alloc_node_mem_map(str
 #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
 static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
 {
-	/*
-	 * We start only with one section of pages, more pages are added as
-	 * needed until the rest of deferred pages are initialized.
-	 */
-	pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
-						pgdat->node_spanned_pages);
 	pgdat->first_deferred_pfn = ULONG_MAX;
 }
 #else
_

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init()
  2018-11-22  9:48 [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init() Wei Yang
  2018-12-20 23:01 ` Andrew Morton
@ 2018-12-20 23:50 ` Alexander Duyck
  2018-12-21  3:39   ` Wei Yang
  1 sibling, 1 reply; 4+ messages in thread
From: Alexander Duyck @ 2018-12-20 23:50 UTC (permalink / raw)
  To: Wei Yang, osalvador; +Cc: akpm, linux-mm

On Thu, 2018-11-22 at 17:48 +0800, Wei Yang wrote:
> When DEFERRED_STRUCT_PAGE_INIT is configured, only the first section of
> each node's highest zone is initialized before defer stage.
> 
> static_init_pgcnt is used to store the number of pages like this:
> 
>     pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
>                                               pgdat->node_spanned_pages);
> 
> because we don't want to overflow zone's range.
> 
> But this is not necessary, since defer_init() is called like this:
> 
>   memmap_init_zone()
>     for pfn in [start_pfn, end_pfn)
>       defer_init(pfn, end_pfn)
> 
> In case (pgdat->node_spanned_pages < PAGES_PER_SECTION), the loop would
> stop before calling defer_init().
> 
> BTW, comparing PAGES_PER_SECTION with node_spanned_pages is not correct,
> since nr_initialised is zone based instead of node based. Even
> node_spanned_pages is bigger than PAGES_PER_SECTION, its highest zone
> would have pages less than PAGES_PER_SECTION.
> 
> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>

This all seems to make sense to me, and appears to be a valid
improvement.

Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>

> ---
>  include/linux/mmzone.h |  2 --
>  mm/page_alloc.c        | 13 ++++++-------
>  2 files changed, 6 insertions(+), 9 deletions(-)
> 
> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
> index ae7a830a21eb..68d7b558924b 100644
> --- a/include/linux/mmzone.h
> +++ b/include/linux/mmzone.h
> @@ -692,8 +692,6 @@ typedef struct pglist_data {
>  	 * is the first PFN that needs to be initialised.
>  	 */
>  	unsigned long first_deferred_pfn;
> -	/* Number of non-deferred pages */
> -	unsigned long static_init_pgcnt;
>  #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
>  
>  #ifdef CONFIG_TRANSPARENT_HUGEPAGE
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 76e12179cd5e..b542d82400cf 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -326,8 +326,13 @@ defer_init(int nid, unsigned long pfn, unsigned long end_pfn)
>  	/* Always populate low zones for address-constrained allocations */
>  	if (end_pfn < pgdat_end_pfn(NODE_DATA(nid)))
>  		return false;
> +
> +	/*
> +	 * We start only with one section of pages, more pages are added as
> +	 * needed until the rest of deferred pages are initialized.
> +	 */
>  	nr_initialised++;
> -	if ((nr_initialised > NODE_DATA(nid)->static_init_pgcnt) &&
> +	if ((nr_initialised > PAGES_PER_SECTION) &&
>  	    (pfn & (PAGES_PER_SECTION - 1)) == 0) {
>  		NODE_DATA(nid)->first_deferred_pfn = pfn;
>  		return true;
> @@ -6451,12 +6456,6 @@ static void __ref alloc_node_mem_map(struct pglist_data *pgdat) { }
>  #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
>  static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
>  {
> -	/*
> -	 * We start only with one section of pages, more pages are added as
> -	 * needed until the rest of deferred pages are initialized.
> -	 */
> -	pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
> -						pgdat->node_spanned_pages);
>  	pgdat->first_deferred_pfn = ULONG_MAX;
>  }
>  #else

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init()
  2018-12-20 23:50 ` Alexander Duyck
@ 2018-12-21  3:39   ` Wei Yang
  0 siblings, 0 replies; 4+ messages in thread
From: Wei Yang @ 2018-12-21  3:39 UTC (permalink / raw)
  To: Alexander Duyck; +Cc: Wei Yang, osalvador, akpm, linux-mm

On Thu, Dec 20, 2018 at 03:50:29PM -0800, Alexander Duyck wrote:
>On Thu, 2018-11-22 at 17:48 +0800, Wei Yang wrote:
>> When DEFERRED_STRUCT_PAGE_INIT is configured, only the first section of
>> each node's highest zone is initialized before defer stage.
>> 
>> static_init_pgcnt is used to store the number of pages like this:
>> 
>>     pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
>>                                               pgdat->node_spanned_pages);
>> 
>> because we don't want to overflow zone's range.
>> 
>> But this is not necessary, since defer_init() is called like this:
>> 
>>   memmap_init_zone()
>>     for pfn in [start_pfn, end_pfn)
>>       defer_init(pfn, end_pfn)
>> 
>> In case (pgdat->node_spanned_pages < PAGES_PER_SECTION), the loop would
>> stop before calling defer_init().
>> 
>> BTW, comparing PAGES_PER_SECTION with node_spanned_pages is not correct,
>> since nr_initialised is zone based instead of node based. Even
>> node_spanned_pages is bigger than PAGES_PER_SECTION, its highest zone
>> would have pages less than PAGES_PER_SECTION.
>> 
>> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
>
>This all seems to make sense to me, and appears to be a valid
>improvement.
>
>Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>

Thanks :-)

>
>> ---
>>  include/linux/mmzone.h |  2 --
>>  mm/page_alloc.c        | 13 ++++++-------
>>  2 files changed, 6 insertions(+), 9 deletions(-)
>> 
>> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
>> index ae7a830a21eb..68d7b558924b 100644
>> --- a/include/linux/mmzone.h
>> +++ b/include/linux/mmzone.h
>> @@ -692,8 +692,6 @@ typedef struct pglist_data {
>>  	 * is the first PFN that needs to be initialised.
>>  	 */
>>  	unsigned long first_deferred_pfn;
>> -	/* Number of non-deferred pages */
>> -	unsigned long static_init_pgcnt;
>>  #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
>>  
>>  #ifdef CONFIG_TRANSPARENT_HUGEPAGE
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index 76e12179cd5e..b542d82400cf 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -326,8 +326,13 @@ defer_init(int nid, unsigned long pfn, unsigned long end_pfn)
>>  	/* Always populate low zones for address-constrained allocations */
>>  	if (end_pfn < pgdat_end_pfn(NODE_DATA(nid)))
>>  		return false;
>> +
>> +	/*
>> +	 * We start only with one section of pages, more pages are added as
>> +	 * needed until the rest of deferred pages are initialized.
>> +	 */
>>  	nr_initialised++;
>> -	if ((nr_initialised > NODE_DATA(nid)->static_init_pgcnt) &&
>> +	if ((nr_initialised > PAGES_PER_SECTION) &&
>>  	    (pfn & (PAGES_PER_SECTION - 1)) == 0) {
>>  		NODE_DATA(nid)->first_deferred_pfn = pfn;
>>  		return true;
>> @@ -6451,12 +6456,6 @@ static void __ref alloc_node_mem_map(struct pglist_data *pgdat) { }
>>  #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
>>  static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
>>  {
>> -	/*
>> -	 * We start only with one section of pages, more pages are added as
>> -	 * needed until the rest of deferred pages are initialized.
>> -	 */
>> -	pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
>> -						pgdat->node_spanned_pages);
>>  	pgdat->first_deferred_pfn = ULONG_MAX;
>>  }
>>  #else

-- 
Wei Yang
Help you, Help me

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2018-12-21  3:39 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-22  9:48 [PATCH] mm: check nr_initialised with PAGES_PER_SECTION directly in defer_init() Wei Yang
2018-12-20 23:01 ` Andrew Morton
2018-12-20 23:50 ` Alexander Duyck
2018-12-21  3:39   ` Wei Yang

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.