mm: Remove is_migrate_highatomic()

[PATCH] mm: Remove is_migrate_highatomic()

Posted by Brendan Jackman 5 months, 2 weeks ago

There are 3 potential reasons for is_migrate_*() helpers:

1. They represent higher-level attributes of migratetypes, like
   is_migrate_movable()

2. They are ifdef'd, like is_migrate_isolate().

3. For consistency with an is_migrate_*_page() helper, also like
   is_migrate_isolate().

It looks like is_migrate_highatomic() was for case 3, but that was
removed in commit e0932b6c1f94 ("mm: page_alloc: consolidate free page
accounting").

So remove the indirection and go back to a simple comparison.

Signed-off-by: Brendan Jackman <jackmanb@google.com>
---
 mm/internal.h   | 5 -----
 mm/page_alloc.c | 2 +-
 2 files changed, 1 insertion(+), 6 deletions(-)

diff --git a/mm/internal.h b/mm/internal.h
index 45b725c3dc030cb22a3cb0ddb262b6300e65ed5a..45da9ff5694f628db18f740718f0a7a06c4548f0 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -1333,11 +1333,6 @@ extern const struct trace_print_flags pageflag_names[];
 extern const struct trace_print_flags vmaflag_names[];
 extern const struct trace_print_flags gfpflag_names[];
 
-static inline bool is_migrate_highatomic(enum migratetype migratetype)
-{
-	return migratetype == MIGRATE_HIGHATOMIC;
-}
-
 void setup_zone_pageset(struct zone *zone);
 
 struct migration_target_control {
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index ca9e6b9633f79ff28448a6008749fa6ca2696332..baead29b3e67bb3ae2291b00d7632aaf87ebcc4c 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -797,7 +797,7 @@ static inline void account_freepages(struct zone *zone, int nr_pages,
 
 	if (is_migrate_cma(migratetype))
 		__mod_zone_page_state(zone, NR_FREE_CMA_PAGES, nr_pages);
-	else if (is_migrate_highatomic(migratetype))
+	else if (migratetype == MIGRATE_HIGHATOMIC)
 		WRITE_ONCE(zone->nr_free_highatomic,
 			   zone->nr_free_highatomic + nr_pages);
 }

---
base-commit: 5654520c361a6638f9d9e889257dfce38c83f653
change-id: 20250821-is-migrate-highatomic-871c0ff05a84

Best regards,
-- 
Brendan Jackman <jackmanb@google.com>

Re: [PATCH] mm: Remove is_migrate_highatomic()

Posted by SeongJae Park 5 months, 2 weeks ago

On Thu, 21 Aug 2025 13:29:47 +0000 Brendan Jackman <jackmanb@google.com> wrote:

> There are 3 potential reasons for is_migrate_*() helpers:
> 
> 1. They represent higher-level attributes of migratetypes, like
>    is_migrate_movable()
> 
> 2. They are ifdef'd, like is_migrate_isolate().
> 
> 3. For consistency with an is_migrate_*_page() helper, also like
>    is_migrate_isolate().
> 
> It looks like is_migrate_highatomic() was for case 3, but that was
> removed in commit e0932b6c1f94 ("mm: page_alloc: consolidate free page
> accounting").
> 
> So remove the indirection and go back to a simple comparison.
> 
> Signed-off-by: Brendan Jackman <jackmanb@google.com>

Acked-by: SeongJae Park <sj@kernel.org>


Thanks,
SJ

[...]

Re: [PATCH] mm: Remove is_migrate_highatomic()

Posted by Lorenzo Stoakes 5 months, 2 weeks ago

On Thu, Aug 21, 2025 at 01:29:47PM +0000, Brendan Jackman wrote:
> There are 3 potential reasons for is_migrate_*() helpers:
>
> 1. They represent higher-level attributes of migratetypes, like
>    is_migrate_movable()
>
> 2. They are ifdef'd, like is_migrate_isolate().
>
> 3. For consistency with an is_migrate_*_page() helper, also like
>    is_migrate_isolate().
>
> It looks like is_migrate_highatomic() was for case 3, but that was
> removed in commit e0932b6c1f94 ("mm: page_alloc: consolidate free page
> accounting").
>
> So remove the indirection and go back to a simple comparison.
>
> Signed-off-by: Brendan Jackman <jackmanb@google.com>

LGTM, so:

Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>

> ---
>  mm/internal.h   | 5 -----
>  mm/page_alloc.c | 2 +-
>  2 files changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/mm/internal.h b/mm/internal.h
> index 45b725c3dc030cb22a3cb0ddb262b6300e65ed5a..45da9ff5694f628db18f740718f0a7a06c4548f0 100644
> --- a/mm/internal.h
> +++ b/mm/internal.h
> @@ -1333,11 +1333,6 @@ extern const struct trace_print_flags pageflag_names[];
>  extern const struct trace_print_flags vmaflag_names[];
>  extern const struct trace_print_flags gfpflag_names[];
>
> -static inline bool is_migrate_highatomic(enum migratetype migratetype)
> -{
> -	return migratetype == MIGRATE_HIGHATOMIC;
> -}
> -
>  void setup_zone_pageset(struct zone *zone);
>
>  struct migration_target_control {
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index ca9e6b9633f79ff28448a6008749fa6ca2696332..baead29b3e67bb3ae2291b00d7632aaf87ebcc4c 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -797,7 +797,7 @@ static inline void account_freepages(struct zone *zone, int nr_pages,
>
>  	if (is_migrate_cma(migratetype))
>  		__mod_zone_page_state(zone, NR_FREE_CMA_PAGES, nr_pages);
> -	else if (is_migrate_highatomic(migratetype))
> +	else if (migratetype == MIGRATE_HIGHATOMIC)
>  		WRITE_ONCE(zone->nr_free_highatomic,
>  			   zone->nr_free_highatomic + nr_pages);
>  }
>
> ---
> base-commit: 5654520c361a6638f9d9e889257dfce38c83f653
> change-id: 20250821-is-migrate-highatomic-871c0ff05a84
>
> Best regards,
> --
> Brendan Jackman <jackmanb@google.com>
>

Re: [PATCH] mm: Remove is_migrate_highatomic()

Posted by Johannes Weiner 5 months, 2 weeks ago

On Thu, Aug 21, 2025 at 01:29:47PM +0000, Brendan Jackman wrote:
> There are 3 potential reasons for is_migrate_*() helpers:
> 
> 1. They represent higher-level attributes of migratetypes, like
>    is_migrate_movable()
> 
> 2. They are ifdef'd, like is_migrate_isolate().
> 
> 3. For consistency with an is_migrate_*_page() helper, also like
>    is_migrate_isolate().
> 
> It looks like is_migrate_highatomic() was for case 3, but that was
> removed in commit e0932b6c1f94 ("mm: page_alloc: consolidate free page
> accounting").
> 
> So remove the indirection and go back to a simple comparison.
> 
> Signed-off-by: Brendan Jackman <jackmanb@google.com>

Oops! Thanks.

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Re: [PATCH] mm: Remove is_migrate_highatomic()

Posted by David Hildenbrand 5 months, 2 weeks ago

On 21.08.25 15:29, Brendan Jackman wrote:
> There are 3 potential reasons for is_migrate_*() helpers:
> 
> 1. They represent higher-level attributes of migratetypes, like
>     is_migrate_movable()
> 
> 2. They are ifdef'd, like is_migrate_isolate().
> 
> 3. For consistency with an is_migrate_*_page() helper, also like
>     is_migrate_isolate().
> 
> It looks like is_migrate_highatomic() was for case 3, but that was
> removed in commit e0932b6c1f94 ("mm: page_alloc: consolidate free page
> accounting").
> 
> So remove the indirection and go back to a simple comparison.
> 
> Signed-off-by: Brendan Jackman <jackmanb@google.com>
> ---

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Cheers

David / dhildenb

Re: [PATCH] mm: Remove is_migrate_highatomic()

Posted by Zi Yan 5 months, 2 weeks ago

On 21 Aug 2025, at 9:29, Brendan Jackman wrote:

> There are 3 potential reasons for is_migrate_*() helpers:
>
> 1. They represent higher-level attributes of migratetypes, like
>    is_migrate_movable()
>
> 2. They are ifdef'd, like is_migrate_isolate().
>
> 3. For consistency with an is_migrate_*_page() helper, also like
>    is_migrate_isolate().
>
> It looks like is_migrate_highatomic() was for case 3, but that was
> removed in commit e0932b6c1f94 ("mm: page_alloc: consolidate free page
> accounting").
>
> So remove the indirection and go back to a simple comparison.
>
> Signed-off-by: Brendan Jackman <jackmanb@google.com>
> ---
>  mm/internal.h   | 5 -----
>  mm/page_alloc.c | 2 +-
>  2 files changed, 1 insertion(+), 6 deletions(-)
>

LGTM. Reviewed-by: Zi Yan <ziy@nvidia.com>

--
Best Regards,
Yan, Zi